Ion Exchange and Solvent Extraction: A Series of Advances, Volume 17 (Ion Exchange and Solvent Extraction Series)

Citation preview

Ion Exchange and Solvent Extraction

Copyright © 2004 by Marcel Dekker, Inc.

Ion Exchange and Solvent Extraction A Series of Advances Volume 17 edited by

Yizhak Marcus

The Hebrew University of Jerusalem Jerusalem, Israel

Arup K.SenGupta

Lehigh University Bethlehem, Pennsylvania, U.S.A.

Jacob A.Marinsky Founding Editor

M ARCEL DEKKER, INC.

Copyright © 2004 by Marcel Dekker, Inc.

NEW YORK • BASEL

Transferred to Digital Printing 2005 Although great care has been taken to provide accurate and current information, neither the author(s) nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage, or liability directly or indirectly caused or alleged to be caused by this book. The material contained herein is not intended to provide specific advice or recommendations for any specific situation. Trademark notice: Product or corporate names may be trademarks or registered trademarks and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress. ISBN: 0-8247-5492-1 Headquarters Marcel Dekker, Inc., 270 Madison Avenue, New York, NY 10016, U.S.A. tel: 212–696–9000; fax: 212–685–4540 Distribution and Customer Service Marcel Dekker, Inc., Cimarron Road, Monticello, New York 12701, U.S.A. tel: 800–228–1160; fax: 845–796–1772 Eastern Hemisphere Distribution Marcel Dekker AG, Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41–61–260–6300; fax: 41–61–260–6333 World Wide Web http://www.dekker.com The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional Marketing at the headquarters address above. Copyright © 2004 by Marcel Dekker, Inc. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Current printing (last digit): 10 9 8 7 6 5 4 3 2 1

Copyright © 2004 by Marcel Dekker, Inc.

Preface

The seventeenth volume of Ion Exchange and Solvent Extraction is concerned with advances in solvent extraction, a field that was also covered in Volume 15. The present book was conceived at ISEC ’02, the International Solvent Extraction Conference held in Capetown, South Africa. Several authors responded to the invitation to contribute comprehensive review papers on subjects showing recent advances and in which they are experts and have published extensively, and other authors were subsequently solicited to do the same. The advances made in such industrial fields as pharmaceutical chemistry and treatment of radioactive wastes— as well as hydrometallurgy—are based on those made in the basic research into new extractants, solvents, and processes. They are also based on the understanding of the mechanisms and kinetics of the extraction and the ability to model them adequately. Finally, advances must also rely heavily on knowledge gained in seemingly unrelated fields, such as solution chemistry, chemical thermodynamics, and chemical engineering. It is expected that some of the advances described in this book will be elaborated on in the forthcoming International Solvent Extraction Conference ISEC ’05, to be held in Beijing, China. Chapter 1, by Perrut, deals with the extraction of mainly pharmaceutical products by means of supercritical fluids, predominantly supercritical carbon dioxide. This “green” solvent extraction technique is very versatile and obviates a stripping operation in order to collect the extracted product. It can be applied on any desired scale, and plant-size applications are described along with advances concerning laboratory-scale extractions. In this context, the application of supercritical fluids in the pharmaceutical industry is described, not only in supercritical fluid extraction (SFE) but also in supercritical fluid fractionation (SFF) and supercritical fluid chromatography (SFC), employed for the purification of natural or synthetic biologically active products. Supercritical fluids are also attractive as reaction media iii Copyright © 2004 by Marcel Dekker, Inc.

iv

Preface

and can be employed with advantage to prepare particles of finely tailored sizes, for both pure compounds and composites, which is useful in drug formulations. Chapter 2, by Bart and Stevens, describes the chemistry and chemical engineering aspects of reactive solvents that include, but are not confined to, what are also called “liquid ion exchangers.” The extraction of zinc ions by di-2-ethylhexyl phosphoric acid has become a recognized industrial test system, and many of the examples given in this chapter deal with this system. Criteria for the choice of solvents and diluents for the processes are described. An extensive discussion follows of the equilibria, distribution constants, and activity coefficients, due to nonidealities in both the aqueous and the organic phases. Definitive instructions or references to the literature on how to apply these activity coefficients are presented and illustrated by the zinc di-2-ethylhexyl phosphoric acid extraction system. The kinetics involved in reactive extraction are then discussed in cases in which only diffusion limits the rate or the controlling step is the rate of the chemical reaction. Experimental set-ups for measuring the kinetics and the fate of the drops in twophase dispersed systems, and how these affect the operation of various types of extraction columns, are described. Chapter 3, by Chiarizia and Herlinger, describes the first of three classes of extractants that have become mature and extensively studied and applied in recent years: symmetrical diesters of alkylene diphosphonic acid. This type of extractant is an example of a liquid cation exchanger that is able to extract selectively many metal ions from acidic solutions. These include alkaline earth metal ions, iron(III), lanthanides, and Am(III), as well as thorium and uranium and tetravalent transuranium element cations. The synthesis of the extractants belonging to this class is described in detail as well as their properties—in particular, their aggregation in diluting solvents, which leads mainly to dimers and hexamers, depending on the extractants and conditions. An important aspect of the use of this class of extractants is the synergism exhibited when suitable compounds and cosolvents are added, which is well exploited in order to effect difficult separations. Finally, returning to the subject of Chapter 1, the use of supercritical fluid extraction—the possibility of employing the diesters of alkylene diphosphonic acid in supercritical carbon dioxide as a “green” diluent—is explored. Chapter 4, by Kolarik, deals with representatives of another class of extractants, namely, solvating extractants: dialkylsulfoxides. These compounds have many properties in common with the well-known organophosphorus extractants, but group with the one. The synthesis and properties of replace the basic these extractants are briefly described, followed by a survey of metal ion extraction systems. These are based mainly on extraction from nitric acid [in which uranium(VI) features prominently as the extracted metal ion], and to a lesser extent on extraction from hydrochloric acid and other media. The dependence of the distribution ratios and rates of extraction on the structure of the extractant is emphasized, and the nature of the extracted species is described in detail for all the systems studied. The selectivities achieved with dialkylsulfoxide extractants are discussed as well as the effects of interfering processes, such as third-phase

Copyright © 2004 by Marcel Dekker, Inc.

Preface

v

formation and radiation damage to the extractant. Finally, enhancement of the utility of the extractants by means of synergism with various kinds of reagents is dealt with in detail. Chapter 5, by Rais and Grüner, concerns a class of extremely hydrophobic and highly acidic extractants: metal bis(dicarbollide)s. The most prominent among them are cobalt bis(dicarbollide) and its chloro-substituted analogs, which are particularly suitable for the extraction of cesium and, when augmented by polyethylene glycols, of strontium. These two elements are among the more troublesome fission products present in nuclear wastes, and efforts to remove them from highly acidic or highly alkaline wastes have previously been stymied by the lack of suitable extractants. The history of the development of the metal bis(dicarbollide)s up to plant-scale use is presented as well as methods for the synthesis of many potentially useful extractants of this class. The selective extraction of cesium and strontium by means of these reagents, alone or in combination with synergists, is described in detail, as are some other applications of these unusual materials. Finally, Chapter 6, by Rais, presents the background for extraction by means of the metal bis(dicarbollide)s and similar hydrophobic anions in terms of general expressions that are valid when entire electrolytes are extracted, as is the case in particular for the alkali metal cations. Extraction curves with polar solvating solvents, such as nitrobenzene or even the dialkylsulfoxides of Chapter 4—where cations are extracted as ion pairs with inorganic anions—often exhibit maxima that are shown to depend on the degree of ionic dissociation of the electrolytes in the organic phase. An algorithm useful for the modeling of such systems is presented. The electrolyte distribution ratios between mutually saturated but waterimmiscible solvents are compared with the standard molar Gibbs energies of transfer of electrolytes between water and the neat solvents and the electrochemical potentials measured with ion-selective electrodes (ISEs) at the interface of two immiscible electrolyte solutions (ITIES), showing the relationships between these useful quantities. This survey of the book’s contents demonstrates the viability of solvent extraction as a separation method used in laboratories and in industry and the continued advances made in finding new extractants and methods for their utilization. Also shown are some advances of the general concepts concerning the mechanisms and rates of extraction that underlie its applications. It is expected that readers of this volume, in conjunction with previous volumes and foreseen future ones, will obtain a stimulating view of the research and application possibilities of solvent extraction. The present volume is the last one edited by Y.M. Since my interest in solvent extraction has waned, I do not feel competent to judge what future significant advances in the field will be so I leave this task to younger persons. Yizhak Marcus Arup K.SenGupta

Copyright © 2004 by Marcel Dekker, Inc.

Contributors to Volume 17

Hans-Jörg Bart Lehrstuhl für Thermische Verfahrenstechnik, Technische Universität Kaiserslautern, Kaiserslautern, Germany R.Chiarizia Chemistry Division, Argonne National Laboratory, Argonne, Illinois, U.S.A. Bohumír Grüner Institute of Inorganic Chemistry, Czech Academy of Sciences, Rez, Czech Republic A.W.Herlinger Department of Chemistry, Loyola University Chicago, Chicago, Illinois, U.S.A. Zdenek Kolarik Consultant, Karlsruhe, Germany Michel Perrut Separex, Champigneulles, France Jirí Rais Nuclear Research Institute Rez plc, Rez Czech Republic Geoffrey W.Stevens Department of Chemical and Biomolecular Engineering, The University of Melbourne, Victoria, Australia

vii Copyright © 2004 by Marcel Dekker, Inc.

Contents

Preface Contributors to Volume 17 Contents of Other Volumes 1.

Applications of Supercritical Fluid Solvents in the Pharmaceutical Industry Michel Perrut I. II. III. IV. V. VI. VII. VIII. IX.

2.

Scope Supercritical Fluid Solvent Properties Applications of SCF as Extraction/Fractionation Solvents Applications of SCF as Chromatography Eluents Applications of SCF as Reaction Media Pollution Abatement Applications of SCF for Particle Design and Drug Formulation Biological Applications Future Trends References

Reactive Solvent Extraction Hans-Jörg Bart and Geoffrey W.Stevens I. Introduction II. Reactive Solvent Extraction Equilibria III. Reactive Mass Transfer

iii vii xiii

1 1 2 7 12 13 14 15 26 27 28 37 37 48 58 ix

Copyright © 2004 by Marcel Dekker, Inc.

x

Contents IV. Solvent Extraction Equipment V. Concluding Remarks References

3.

Symmetrical P,P’-Disubstituted Esters of Alkylenediphosphonic Acids as Reagents for Metal Solvent Extraction 85 R.Chiarizia and A.W.Herlinger I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII.

4.

5.

68 76 77

Introduction Synthesis Spectroscopic Studies Aggregation Studies Solvent Extraction of Metal Ions at Low Loading Solvent Extraction of Metal Ions at High Loading Enthalpy and Entropy Changes in Metal Solvent Extraction Intra-Lanthanide Ion Separations Synergistic Extraction of Metal Ions Extraction Chromatographic Applications Reagents for Use in Supercritical Fluid Extraction Conclusions Nomenclature References

85 87 94 103 110 120 126 132 138 145 149 155 157 158

Sulfoxide Extractants and Synergists Zdenek Kolarik

165

I. II. III. IV. V. VI. VII.

165 169 195 209 213 217 221 233 234 236

Introduction Extraction from Nitrate Media Extraction from Chloride Media Extraction from Other Media Selectivity of the Extraction Interfering Phenomena Synergism Nomenclature Appendix: Physical Properties of Sulfoxides References

Extraction with Metal Bis(dicarbollide) Anions: Metal Bis(dicarbollide) Extractants and Their Applications in Separation Chemistry 243 Jirí Rais and Bohumír Grüner I. Introduction 243 II. Synthesis and Properties of Metal Bis(dicarbollide)s and Other Cluster Boron Compounds Aimed for Extraction Purposes 247

Copyright © 2004 by Marcel Dekker, Inc.

Contents III. Extraction with Basic and Halogenated Cobalt Bis(dicarbollide)s IV. Extraction with Cobalt Bis(dicarbollide)s and Synergists V. Extraction with Functionalized Metal Bis(dicarbollide)s and Other Boron Extractants VI. Chloro-Protected Bis(dicarbollide) Technologies for Extraction of Fission Products and Actinide Cations from Radioactive Wastes VII. Analytical and Other Applications of Extraction Systems with Metal Bis(dicarbollide)s VIII. Conclusions Supplementary Material Symbols References 6.

Principles of Extraction of Electrolytes Jiri Rais I. II. III. IV

Introduction Aims and Scope of the Review Description of the Systems Characteristic Examples of Equilibria in Extraction of Electrolytes V Gibbs Energies of Transfer and Some Semiempirical Models VI. Conclusions Symbols References

Copyright © 2004 by Marcel Dekker, Inc.

xi

282 283 299

306 312 322 322 323 324 335 335 338 338 343 357 378 378 381

Contents of Other Volumes

Volumes 1–4, 6 out of print Volume 5 NEW INORGANIC ION EXCHANGERS A.Clearfield, G.H.Nancollas, and R.H.Blessing APPLICATION OF ION EXCHANGE TO ELEMENT SEPARATION AND ANALYSIS F.W.E.Strelow PELLICULAR ION EXCHANGE RESINS IN CHROMATOGRAPHY Csaba Horvath Volume 7 INTERPHASE MASS TRANSFER RATES OF CHEMICAL REACTIONS WITH CROSSLINKED POLYSTYRENE Gabriella Schmuckler and Shimon Goldstein INFLUENCE OF POLYMERIC MATRIX STRUCTURE ON PERFORMANCE OF ION-EXCHANGE RESINS V.A.Davankov, S.V.Rogozhin, and M.P.Tsyurupa xiii Copyright © 2004 by Marcel Dekker, Inc.

xiv

Contents of Other Volumes

SPECTROSCOPIC STUDIES OF ION EXCHANGERS Carla Heitner-Wirguin ION-EXCHANGE MATERIALS IN NATURAL WATER SYSTEMS Michael M.Reddy THE THERMAL REGENERATION OF ION-EXCHANGE RESINS B.A.Bolto and D.E.Weiss Volume 8 METAL EXTRACTION WITH HYDROXYOXIMES Richard J.Whewell and Carl Hanson ELECTRICAL PHENOMENA IN SOLVENT EXTRACTION Giancarlo Scibona, Pier Roberto Dansei, and Claudio Fabiani EXTRACTION WITH SOLVENT-IMPREGNATED RESINS Abraham Warshawsky SOLVENT EXTRACTION OF ELEMENTS OF THE PLATINUM GROUP Lev M.Gindin SOLVENT EXTRACTION FROM AQUEOUS-ORGANIC MEDIA Jiri Hala Volume 9 ION-EXCHANGE PROCESSES USED IN THE PRODUCTION OF ULTRAPURE WATER REQUIRED IN FOSSIL FUEL POWER PLANTS Calvin Calmon A SYSTEMATIC APPROACH TO REACTIVE ION EXCHANGE Gilbert E.Janauer, Robert E.Gibbons, Jr., and William E.Bernier ION-EXCHANGE KINETICS IN SELECTIVE SYSTEMS Lorenzo Liberti and Roberto Passino SORPTION AND CHROMATOGRAPHY OF ORGANIC IONS G.V.Samsonov and G.E.Elkin

Copyright © 2004 by Marcel Dekker, Inc.

Contents of Other Volumes

xv

THERMODYNAMICS OF WATER SORPTION OF DOWEX 1 OF DIFFERENT CROSSLINKING AND IONIC FORM Zoya I.Sosinovich, Larissa V.Novitskaya, Vladimir S.Soldatov, and Erik Högfeldt DOUBLE-LAYER IONIC ADSORPTION AND EXCHANGE ON POROUS POLYMERS Frederick F.Cantwell HUMIC-TRACE METAL ION EQUILIBRIA IN NATURAL WATERS Donald S.Gamble, Jacob A.Marinsky, and Cooper H.Langford Volume 10 SOLVENT EXTRACTION OF INDUSTRIAL ORGANIC SUBSTANCES FROM AQUEOUS STREAMS C.Judson King and John J.Senetar LIQUID MEMBRANES Richard D.Noble, J.Douglas Way, and Annett L.Bunge MIXED SOLVENTS IN GAS EXTRACTION AND RELATED PROCESSES Gerd Brunner INTERFACIAL PHENOMENA IN SOLVENT EXTRACTION Valery V.Tarasov and Gennady A.Yagodin SYNERGIC EXTRACTIONS OF ZIRCONIUM (IV) AND HAFNIUM (IV) Jiri Hala Volume 11 CHEMICAL THERMODYNAMICS OF CATION EXCHANGE REACTIONS: THEORETICAL AND PRACTICAL CONSIDERATIONS Steven A.Grant and Philip Fletcher A THREE-PARAMETER MODEL FOR SUMMARIZING DATA IN ION EXCHANGE Erik Högfeldt DESCRIPTION OF ION-EXCHANGE EQUILIBRIA BY MEANS OF THE SURFACE COMPLEXATION THEORY Wolfgang H.Höll, Matthias Franzreb, Jürgen Horst, and Siefried H.Eberle

Copyright © 2004 by Marcel Dekker, Inc.

xvi

Contents of Other Volumes

SURFACE COMPLEXATION OF METALS BY NATURAL COLLOIDS Garrison Sposito A GIBBS-DONNAN-BASED ANALYSIS OF ION-EXCHANGE AND RELATED PHENOMENA Jacob A.Marinsky INFLUENCE OF HUMIC SUBSTANCES ON THE UPTAKE OF METAL IONS BY NATURALLY OCCURING MATERIALS James H.Ephraim and Bert Allard Volume 12 HIGH-PRESSURE ION-EXCHANGE SEPARATION IN RARE EARTHS Liquan Chen, Wenda Xin, Changfa Dong, Wangsuo Wu, and Sujun Yue ION EXCHANGE IN COUNTERCURRENT COLUMNS Vladimir I. Gorshkov RECOVERY OF VALUABLE MINERAL COMPONENTS FROM SEAWATER BY ION-EXCHANGE AND SORPTION METHODS Ruslan Khamizov, Dmitri N.Muraviev, and Abraham Warshawsky INVESTIGATION OF INTRAPARTICLE ION-EXCHANGE KINETICS IN SELECTIVE SYSTEMS A.I.Kalinitchev EQUILIBRIUM ANALYSIS OF COMPLEXATION IN ION EXCHANGERS USING SPECTROSCOPIC AND DISTRIBUTION METHODS Hirohiko Waki ION-EXCHANGE KINETICS IN HETEROGENEOUS SYSTEMS K.Bunzl EVALUATION OF THE ELECTROSTATIC EFFECT ON METAL ION-BINDING EQUILIBRIA IN NEGATIVELY CHARGED POLYION SYSTEMS Tohru Miyajima ION-EXCHANGE EQUILIBRIA OF AMINO ACIDS Zuyi Tao ION-EXCHANGE SELECTIVITIES OF INORGANIC ION EXCHANGERS Mitsuo Abe

Copyright © 2004 by Marcel Dekker, Inc.

Contents of Other Volumes

xvii

Volume 13 EXTRACTION OF SALTS BY MIXED LIQUID ION EXCHANGERS Gabriella Schmuckler and Gideon Harel ACID EXTRACTION BY ACID-BASE-COUPLED EXTRACTANTS Aharon M.Eyal HOST-GUEST COMPLEXATION AS A TOOL FOR SOLVENT EXTRACTION AND MEMBRANE TRANSPORT OF (BIO)ORGANIC COMPOUNDS Igor V.Pletnev and Yuri A.Zolotov NEW TECHNOLOGIES FOR METAL ION SEPARATIONS: POLYETHYLENE GLYCOL BASED-AQUEOUS BIPHASIC SYSTEMS AND AQUEOUS BIPHASIC EXTRACTION CHROMATOGRAPHY Robin D.Rogers and Jianhua Zhang DEVELOPMENTS IN SOLID-LIQUID EXTRACTION BY SOLVENTIMPREGNATED RESINS José Luis Cortina and Abraham Warshawsky PRINCIPLES OF SOLVENT EXTRACTION OF ALKALI METAL IONS: UNDERSTANDING FACTORS LEADING TO CESIUM SELECTIVITY IN EXTRACTION BY SOLVATION Bruce A.Moyer and Yunfu Sun

Volume 14 POLYMER-SUPPORTED REAGENTS: THE ROLE OF BIFUNCTIONALITY IN THE DESIGN OF ION-SELECTIVE COMPLEXANTS Spiro D.Alexandratos RECOVERY OF VALUABLE SPECIES FROM DISSOLVING SOLIDS USING ION EXCHANGE Jannie S.J.van Deventer, P.G.R.de Villiers, and L.Lorenzen POLYMERIC LIGAND-BASED FUNCTIONALIZED MATERIALS AND MEMBRANES FOR ION EXCHANGE Stephen M.C Ritchie and Dibakar Bhattacharyya BIOSORPTION OF METAL CATIONS AND ANIONS Bohumil Volesky, Jinbai Yang, and Hui Niu

Copyright © 2004 by Marcel Dekker, Inc.

xviii

Contents of Other Volumes

SYNTHESIS AND APPLICATION OF FUNCTIONALIZED ORGANOCERAMIC SELECTIVE ADSORBENTS Lawrence L.Tavlarides and J.S.Lee ENVIRONMENTAL SEPARATION THROUGH POLYMERIC LIGAND EXCHANGE Arup K.SenGupta IMPRINTED METAL-SELECTIVE ION EXCHANGER Masahiro Goto SYNTHESIS AND CHARACTERIZATION OF A NEW CLASS OF HYBRID INORGANIC SORBENTS FOR HEAVY METALS REMOVAL Arthur D.Kney and Arup K.SenGupta

Volume 15 AN INTEGRATED METHOD FOR DEVELOPMENT AND SCALING UP OF EXTRACTION PROCESSES Baruch Grinbaum DESIGN OF PULSED EXTRACTION COLUMNS Alfons Vogelpohl and Hartmut Haverland PURIFICATION OF NICKEL BY SOLVENT EXTRACTION Kathryn C.Sole and Peter M.Cole TREATMENT OF SOILS AND SLUDGES BY SOLVENT EXTRACTION IN THE UNITED STATES Richard J.Ayen and James D.Navratil THE DESIGN OF SOLVENTS FOR LIQUID-LIQUID EXTRACTION Braam van Dyk and Izak Nieuwoudt EXTRACTION TECHNOLOGY FOR THE SEPARATION OF OPTICAL ISOMERS André B.de Haan and Béla Simándi REGULARITIES OF EXTRACTION IN SYSTEMS ON THE BASIS OF POLAR ORGANIC SOLVENTS AND USE OF SUCH SYSTEMS FOR SEPARATION OF IMPORTANT HYDROPHOBIC SUBSTANCES Sergey M.Leschev DEVELOPMENTS IN DISPERSION-FREE MEMBRANE-BASED EXTRACTION-SEPARATION PROCESSES Anil Kumar Pabby and Ana-Maria Sastre

Copyright © 2004 by Marcel Dekker, Inc.

Contents of Other Volumes

xix

Volume 16 ADSORPTION AND ION-EXCHANGE PROPERTIES OF ENGINEERED ACTIVATED CARBONS AND CARBONACEOUS MATERIALS Michael Streat, Danish J.Malik, and Basudeb Saha ENTROPY-DRIVEN SELECTIVE ION EXCHANGE FOR HYDROPHOBIC IONIZABLE ORGANIC COMPOUNDS (HIOCs) Ping Li and Arup K.SenGupta ION-EXCHANGE ISOTHERMAL SUPERSATURATION: CONCEPT, PROBLEMS, AND APPLICATIONS Dmitri N.Muraviev and Ruslan Khamizov METAL SEPARATION BY pH-DRIVEN PARAMETRIC PUMPING Wolfgang H.Höll, Randolf Kiefer, Cornelia Stöhr, and Christian Bartosch SELECTIVITY CONSIDERATIONS IN MODELING THE TREATMENT OF PERCHLORATE USING ION-EXCHANGE PROCESSES Anthony R.Tripp and Dennis A.Clifford ION-EXCHANGE KINETICS FOR ULTRAPURE WATER Dennis F.Hussey and Gary L.Foutch

Copyright © 2004 by Marcel Dekker, Inc.

1 Applications of Supercritical Fluid Solvents in the Pharmaceutical Industry Michel Perrut Separex, Champigneulles, France

I. SCOPE At the beginning of this new millennium, the pharmaceutical industry is facing many challenges in relation to the rapid and endless growth and aging of the world population, and a growing gap between health care systems in industrialized and developing countries. While higher and higher quality standards are required by authorities and consumers, and massive low-cost supply of drugs is strongly demanded by developing countries, it appears more and more difficult to introduce innovative new drugs and to improve the therapeutic efficacy against numerous pathologies. Furthermore, the industry must also make a continuous effort to move to environmentally friendly processes. Regarding the environmental protection, one of the first requirements of the industry is to move to “green chemistry” and to avoid potentially harmful solvents. As we will try to demonstrate in this chapter, the use of supercritical fluid (SCF) solvents is a promising route to both reducing pollutant release and improving the final drug quality and efficacy through innovative processes for active substance preparation and drug formulation. In fact, most applications of supercritical fluid solvents are based on the use of carbon dioxide, pure or with ethanol added, as it presents the definitive advantages, being a “green,” abundant, and cheap solvent perfectly adequate to process food 1 Copyright © 2004 by Marcel Dekker, Inc.

2

Perrut

or pharmaceutical products at a temperature near to ambient. Many processes are now under development: • • •

Supercritical fluid extraction (SFE), supercritical fluid fractionation (SFF), and supercritical fluid chromatography (SFC), for extraction and purification of natural or synthetic active products. Supercritical fluids as reaction media. Supercritical fluid drug formulation by manufacturing innovative therapeutic particles, either of pure active compounds or composites of excipient and active compounds.

As thousands of publications can be found in the domain, this review does not present an exhaustive survey, but aims to guide the scientists to consider supercritical fluid solvents as a new tool to be envisaged to open innovative routes to solve their problems.

II. SUPERCRITICAL FLUID SOLVENT PROPERTIES Pure compounds can be found in three states: solid, liquid, and vapor or gas. On the (pressure, temperature) diagram as presented in Fig. 1, the three regions corresponding to these three states are separated by curves that meet at the triple point. Surprisingly, the vaporization/liquefaction curve presents an end point called critical point (Pc, Tc). Beyond this point (P>Pc and T>Tc), only one phase exists, called supercritical fluid (SCF). At the critical point itself, the fluid compressibility

Figure 1 General pressure-temperature diagram for pure compounds.

Copyright © 2004 by Marcel Dekker, Inc.

SCF Solvents in the Pharmaceutical Industry

3

becomes infinite, meaning that the fluid density rapidly varies with a slight change in pressure at constant temperature. Moreover, even out of the critical region itself, an SCF exhibits large changes in density—and consequently in solvent power— and other physicochemical properties with pressure or temperature. These may be roughly considered as intermediate between those of a low-viscosity liquid and a compressed gas (see Table 1). Similarly, when the compound is maintained at a pressure above its critical pressure and at a temperature below its critical temperature, it is called subcritical liquid (P>Pc and TPc and T>Tc) and to a liquefied gas (PH2DEH[EDP]>H2DEH[BuDP]. Solvent extraction and aggregation studies confirm this order. Stripping of metal ions from the extractant, for example, is difficult with H2DEH[MDP], whereas for H2DEH[EDP] and H2DEH[BuDP] with weaker binding metal stripping is more readily accomplished. In the systems studied, the frequency of νasym(POO-) is especially sensitive to the nature of the metal ion while the frequency of νsym(POO-) remains relatively

Figure 2 Ionic potential Φ of the metal ion vs. ∆ν=[νasym(POO-)-νsym(POO-)] for 2-ethylhexyl diesters of alkylenediphosphonic acids (circles: H2DEH[MDP], squares: H2DEH[EDP], triangles: H2DEH[BuDP]). (Adapted in part from Ref. 68.)

Copyright © 2004 by Marcel Dekker, Inc.

102

Chiarizia and Herlinger

constant. The highest νasym(POO-) values are observed in the sodium salts while the lowest energy values occur in the iron compounds. The decrease in νasym (POO-) and consequent decrease in ∆ν with the charge to radius ratio of the metal ion are a result of the increased strength of the metal-phosphonate interaction. The interaction is weakest in the sodium salts and becomes progressively stronger in the other studied compounds. Proceeding from the sodium salts to compounds of metal ions with very high ionic potentials, increased P–O bond polarization is expected to occur. In compounds containing very highly charged metal ions, such as Th(IV) and Fe(III), bond polarization might occur to such a great extent that the resulting metal-oxygen bond has substantial covalent character. Nature of the Metal-Diphosphonate Interaction. Evidence of the covalent character of the Fe(III)-diphosphonate interaction is provided by the far-infrared spectra for the sodium, copper, selected lanthanide, and iron(III) salts of H2DEH[MDP] [69]. In the region below 600 cm-1 where the POO- and POC deformation modes appear, the Fe(III)-H2DEH[MDP] complex has an additional strong absorption band at 256 cm-1 that is sensitive to the mass of the Fe isotope used and is absent in the spectrum of the free acid and the other compounds investigated. This band and another band at 551 cm-1, which is also sensitive to the mass of the Fe isotope, were assigned as Fe–O stretching vibrations based on their frequency and 54Fe isotopic shift compared with related iron(III) compounds [69]. The appearance of Fe–O stretching bands in the anhydrous iron-H2DEH[MDP] complex indicates that the Fe-diphosphonate interaction has a substantial covalent component. A metal-oxygen (M–O) stretching mode will be infrared active only if the metal-oxygen bond is sufficiently covalent. The absence of M–O stretching bands in the lanthanide compounds indicates that in these salts the binding is predominantly ionic. Based on these findings, it is evident that the changes observed in the P–O bond order in metal-diphosphonate compounds arise from a combination of covalent and bond polarization effects. Far-infrared results for the H2DEH[EDP]- and H2DEH[BuDP]-iron complexes show that alkylene bridge length affects the frequency of the iron-oxygen stretching modes [68]. The frequencies of the Fe–O bands decrease as the length of the alkylene bridge increases, with the lower energy band appearing at 254 and 249 cm-1 for H2DEH[EDP] and H2DEH[BuDP], respectively. The lower Fe–O stretching frequencies indicate weaker iron-oxygen binding, consistent with the conclusions based on the ∆ν values discussed above. Since the value of ∆ν depends upon the ionic potential of the metal ion, it may be used to obtain qualitative information about the nature of the M–O interaction. When both oxygen atoms of the anion of an organophosphorus extractant interact with the metal ion, high ∆ν values are indicative of high ionic character in M–O bonds while low ∆ν values indicate a high degree of covalent character [68].

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

103

IV. AGGREGATION STUDIES When alkylenediphosphonic acid diesters are dissolved in nonpolar diluents, selfassembly into more complex structures is suggested by the very structure of this class of compounds. In fact, the presence of both hydrogen bond donors (–OH) and acceptors(=O) in structure I clearly indicates that hydrogen bonding is the primary factor in driving the aggregation process. This is confirmed in the infrared spectra by the frequency shifts observed for the phosphoryl bands as compared to those of the tetraesters. Band shifts to lower energy are spectral features characteristic of systems with strong intermolecular hydrogen bonding. Recent osmometric measurements on tetracyclohexyl and other tetraalkyl esters of methylenediphosphonic acid have unequivocally demonstrated that when all the -OH groups in Structure I are esterified, no aggregation takes place and the tetraesters exist in solution as monomers [26], confirming that strong hydrogen bonding is key to aggregate stability. Information about the specific types of aggregates formed through hydrogen bonding (i.e., dimers vs. higher aggregates) cannot be provided by the infrared spectra. Useful indications, however, can be deduced from the behavior of monofunctional organophosphorous acidic extractants, such as di(2-ethylhexyl) phosphoric acid (HDEHP) and mono(2-ethylhexyl) phosphoric acid (H2MEHP), which are known to strongly aggregate in nonpolar diluents [72, 73]. Monoprotic acids usually dimerize to form an ring, analogous to that formed in the familiar dimerization of carboxylic acids. , in the Etter hydrogen bond assembly classification [74], denotes an 8-membered ring structure containing two hydrogen bond donors and two hydrogen bond acceptors. Hydrogen bonding in organophosphorus acid dimers, however, is known to be stronger than in carboxylic acid dimers [75, 76]. The aggregation behavior of diprotic acids such as H2MEHP is more complicated. When two monomers hydrogen bond to form a dimeric species containing one ring, the resulting aggregate has additional –OH groups that can serve as sites for further aggregation [77]. This is shown schematically in Structure II:

Structure II

This scheme rationalizes why HDEHP is dimeric in benzene [73, 78], while H2MEHP can have an aggregation number as high as 12 in the same diluent [78, 79]. Since HDEHP and H2MEHP can be regarded as monofunctional analogues of

Copyright © 2004 by Marcel Dekker, Inc.

104

Chiarizia and Herlinger

the diphosphonic acids shown in Structure I, an interesting question arose in our investigation. That is, would the aggregation behavior of the diphosphonic acid diesters be similar to that of diprotic acids, or would the aggregation be limited to the formation of dimers as is the case for monoprotic acids? Our physicochemical data revealed that both behaviors are possible, with the value of n in Structure I, i.e., the number of carbon atoms in the alkylene bridge connecting the phosphorus atoms, as the determining factor [37, 38, 41, 65, 80–83].

A. Vapor Pressure Osmometry (VPO) Measurements Vapor pressure osmometry (VPO) can be used to measure the aggregation of an extractant in an organic solution by comparing its behavior to that of a monomeric standard [84]. In our studies, experimental data were used to obtain plots of osmometer response (the electrical potential µV needed to equilibrate an electric circuit comprising two thermistors, one wet by the pure solvent, the other by the test solution) vs. solute concentration. The experimental details can be found in the original publications [37, 38, 41, 65, 82, 83]. Figure 3 collects VPO data obtained for diphosphonic acid diesters in toluene, a convenient diluent for this type of investigations. Panel A shows the data for the 2-ethylhexyl diesters, while the data for the 3-trimethylsilylpropyl diesters are presented in panel B. At each solute concentration, the electrical potential values for the diphosphonic acid diesters are considerably lower than the value for the monomeric standard. Since vapor pressure is a colligative property, its lowering is proportional to the number of particles in solution and thus, consequently, related to the aggregation of the solute. In the simplest case, when the electrical potential vs. concentration data are linear, the number-average aggregation number nav is given directly by the ratio of the slopes of the monomeric standard and the extractant solution. As can be seen in Figure 3 for H 2 DEH[MDP], H 2 DTMSP[MDP], H2DTMSP[PrDP] and H2DTMSP[PDP], compounds for which n in Structure 1 is odd (1, 3, or 5), the value of nav is very close to 2, indicating that these extractants exist in toluene as dimers. Thus, in terms of aggregation behavior, independent of the esterifying groups, the methylene-, propylene- and pentylenediphosphonic acid diesters behave similar to the monoprotic monofunctional analogues in Structure II. Since the VPO data are linear, values for the dimerization constants cannot be estimated. However, it is evident that the dimerization constants are sufficiently large that the compounds are completely dimerized over the entire concentration range investigated. The compounds for which n in Structure 1 is an even number (2, 4, or 6), form larger aggregates, analogous to the diprotic monofunctional analogues in Structure II. For example, the slopes of the VPO data for H2DEH[BuDP] and H2DEH[EDP] (panel a) indicate that these diesters form primarily trimeric and hexameric aggregates, respectively. Similar behavior is exhibited by the silyl-containing diesters (panel b). The nav values for H2DEH[BuDP] and H2DTMSP[EDP] shown

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

105

Figure 3 VPO data for diphosphonic acid diesters in toluene at 25°C. (a) Monomeric standard: bibenzyl (squares); data for H2DEH[MDP] (circles), H2DEH[EDP] (diamonds) and H2DEH[BuDP] (triangles). (b) Monomeric standard: sucrose octaacetate (full squares); data for H2DTMSP[MDP] (full circle), H2DTMSP[EDP] (empty circles), H2DTMSP[PrDP] (up triangles), H 2DTMSP[BuDP] (diamonds), H 2 DTMSP [PDP](empty squares), and H2DTMSP[HDP] (down triangles).

Copyright © 2004 by Marcel Dekker, Inc.

106

Chiarizia and Herlinger

in Fig. 3 (3.3 and 5.5, respectively) are those obtained from a best straight-line fit of the data. Closer inspection reveals that these data are not linear, indicating an equilibrium between two or more species with different aggregation numbers. This behavior is particularly evident for the H2DTMSP[BuDP] and H2DTMSP[HDP] diesters (panel b). In this case, the VPO data can be used to calculate values for the aggregation constants following the procedure described earlier [85]. The values of the aggregations constants of the various diesters investigated are summarized in Table 5. It appears from Table 5 that the aggregation of the diphosphonic acid diesters exhibits an even-odd effect as the number of the methylene bridging groups varies. This effect is likely to be due to the “zig-zag” pattern adopted by the alkylene chain separating the phosphorus atoms. This pattern controls the orientation of the and POH groups and changes the geometry of the hydrogen bonded aggregates that can be formed. This even-odd effect also manifests itself in the melting points of the partially esterified diphosphonic acids [40] and the parent acids [18], with the molecules containing an even number of bridging methylene groups exhibiting higher than expected melting points. Table 5 shows a significant difference between the 2-ethyhexyl (EH) and corresponding 3-trimethylsilyl-1-propyl (TMSP) diesters, the latter being less strongly aggregated than the former, due to larger steric hindrance by the TMSP groups. Also, within the series of TMSP diesters, the stability of the hexameric aggregate decreases with an increase in the number of bridging methylene groups, indicating that self-assembly into hexamers is more difficult for molecules with longer alkylene chains. This phenomenon has practical consequences in metal ion extraction studies. Speciation diagrams [37, 41, 83] have shown that, under the Table 5 Aggregation Constants of Alkylenediphosphonic Acids Diesters in Toluene

a b

Equilibrium constant too large to be calculated from VPO data; Calculated from metal distribution data (see Section V).

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

107

conditions typically used for solvent extraction studies, the presence of monomeric species cannot be ignored for weakly aggregated diesters. The presence of monomeric extractants must be taken into account to explain some solvent extraction data.

B. SANS Measurements Light, X-ray, and neutron scattering are powerful techniques in structural studies of polymers, micellar aggregates, and other materials [86]. The nanoscale size of the aggregates and the large amount of hydrogen atoms in organic materials make small-angle neutron scattering (SANS) particularly well suited for investigating extractant aggregates in organic diluents [87]. The SANS technique is based on the large difference in the neutron scattering properties of hydrogen and deuterium atoms. Dissolution of an extractant in a deuterated diluent provides the neutron scattering contrast required to make the solute particles “visible” against the solvent background. SANS measurements of deuterated toluene solutions of H 2DEH[MDP], H2DEH[EDP], H2DEH[BuDP] and their metal complexes were made using the time-of-flight small-angle neutron diffractometers (SAD and SAND) at the Intense Pulsed Neutron Source (IPNS) at ANL. The characteristics of the diffractometers, the background corrections and the procedure for placing the data on an absolute scale can be found in the original publications and references therein [80, 81, 88]. The SANS data were obtained as plots of scattering intensity I (cm-1) vs. momentum transfer, Q(Å-1) (Q=(4π/␭) sin θ, where θ is half the scattering angle and λ is the wavelength of the probing neutrons). The SANS scattering signals were interpreted using the Guinier analysis {ln[I(Q)] vs. Q2} [89]. The Guinier fit was used to determine the molecular weight of the extractant aggregates and, hence, the weightaverage aggregation number nw. The SANS results obtained with metal loaded extractants are discussed later in this chapter. Here we summarize the results obtained from SANS measurements on deuterotoluene solutions of the extractants alone. Figure 4 shows the SANS data for 0.1 M solutions of H2DEH[MDP], H2DEH[EDP] and H2DEH[BuDP] (panel A), and the Guinier fits of the data (panel B). The extrapolation to Q2=0 of the Guinier straight lines shown in panel B provided the I(0) values which were used to calculate the aggregation number of the extractants [80, 81]. The results of these calculations confirmed that H2DEH[MDP] exists in solution as a dimer, H2DEH[EDP] as a hexamer, and H2DEH[BuDP] predominantly as a trimer.

C. Structure of the Aggregates The MDP, PrDP, and PDP dimers could have a variety of different structures. Some possibilities, based on various combinations of intra- and intermolecular hydrogen bonds, are shown in Structures III–VI below for the MDP dimer.

Copyright © 2004 by Marcel Dekker, Inc.

108

Chiarizia and Herlinger

Figure 4 SANS behavior for 0.1 M H2DEH[MDP], H2DEH[EDP], and H2DEH [BuDP] in deuterotoluene. (a) Log-log plots of scattering data; (b) Guinier fits of data.

The results from continuous variation infrared spectroscopic studies, in which the diluent in a solution-containing H2DEH[MDP] was progressively replaced with the depolymerizing diluent 1-decanol, were consistent with Structures V and VI [66]. To fully elucidate the relative stabilities of Structures III–VI, molecular mechanics calculations were performed. A variety of starting structures containing two (such as III and IV) or four (such as V and VI) intermolecular hydrogen bonds were generated and geometrically optimized. For ease of computation, the 2-ethylhexyl groups in

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

109

H2DEH[MDP] were replaced by methyl groups. Two highly hydrogen-bonded dimer structures were found to be the most stable. They exhibited the hydrogenbonding patterns of Structures V and VI. The lowest energy conformation found for Structure VI was slightly higher in energy than the lowest energy conformation found for Structure V [66].

Structure III

Structure V

Structure IV

Structure VI

The predominantly trimeric aggregates of the BuDP diesters, and the hexameric aggregates of the EDP and HDP diesters quite possibly have a cyclic structure, based on spectroscopic evidence indicating that all the hydrogen bonds in these aggregates are equivalent [37, 38]. In our attempts to obtain more detailed information about the three-dimensional shape of the H2DEH[EDP] aggregate, we were able to fit the SANS data for H2DEH[EDP] using the equation for a homogeneous sphere [81, 90]. The best fit provided a radius R of the spherical aggregate equal to 11.8±0.2 Å. The hexamer probably assumes a spherical shape best illustrated by the seams on a tennis ball. A representation of this shape for the H2DEH[EDP] hexamer is shown in Structure VII. In this aggregate, the alkyl groups are likely oriented outward (toward the lipophilic solvent), and a large hydrophilic internal cavity is available to accommodate metal cations and/or water molecules. Thus, Structure VII strongly resembles a reverse micelle. Although no SANS measurements have been performed on the TMSP diesters, it is very likely that the structures of their aggregates parallel those of the 2-ethylhexyl analogues.

Copyright © 2004 by Marcel Dekker, Inc.

110

Chiarizia and Herlinger

Structure VII

V.

SOLVENT EXTRACTION OF METAL IONS AT LOW LOADING

The preceding sections were propaedeutic for the interpretation of data on metal ion extraction by alkylenediphosphonic acid diesters dissolved in water-immiscible diluents. Spectroscopic and aggregation studies have shown that a number of effects arise when the distance between the phosphorus atoms of a diester is increased by increasing n in Structure I. All these effects, in turn, have an impact on metal extraction behavior. First, the basicity of the group increases and is expected to follow the order with the acidity of the most acidic POH proton following the opposite order (see Section III). The increase in basicity of the donor group increases the strength of the interaction with positively charged ions and thus enhances metal ion extraction. On the other hand, the concomitant decrease in acid strength increases the competition between the proton and the metal ion for the ligand, thus opposing metal extraction.

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

111

The second effect concerns the cooperative bonding of the metal ion by both phosphonate groups of the ligand, which strongly enhances metal extraction. While this behavior is expected for the methylenediphosphonic acid diesters (n=1 in Structure I), cooperative binding becomes less likely for ligands with n>1. When the separation between the phosphonate groups is sufficiently large, the solvent extraction behavior of the diesters should parallel that of the analogous monofunctional organophosphorus extractants. The third effect is due to the size of the chelate ring formed when both phosphonate groups bind to the same metal cation. The methylenediphosphonic acid diesters can form highly stable six-membered rings, as can be seen in Structure I. For ligands with n≥2, larger and progressively less stable chelate rings would be formed, resulting in greatly diminished extraction efficiency. Finally, the effect of extractant aggregation on metal extraction chemistry manifests itself in several ways. Aggregation shifts the partition equilibrium of the extractant toward the organic phase reducing its concentration in the aqueous phase. Further, aggregation of the metal-extractant complexes shifts the overall extraction equilibria to the right. Both effects increase the metal distribution ratio. The presence of large extractant aggregates in the organic phase also affects the stoichiometry of the extraction reaction. This can be easily understood by considering the following simplified equilibria describing the extraction of a metal ion Mz+ by a highly aggregated acidic extractant HA: (9) where the bar denotes organic phase species. The metal distribution ratio, defined as the ratio of the equilibrium metal concentrations in the organic and aqueous phase, can be written as (10) where K is the equilibrium constant for the extraction reaction. Assuming that the concentration of monomeric extractant is negligible, it follows that (11) The analytical concentration of the extractant, C HA, is proportional to the concentration of the extractant aggregate with the proportionality constant equal to the aggregation number n. Substituting Eq. (11) in Eq. (10), one obtains (12) which demonstrates that in a log-log plot of D vs. CHA, the slope of the extractant dependency will be unity regardless of the charge on the cation. This type of behavior has been reported for the extraction of actinide and lanthanide ions by H2MEHP in aromatic diluents [77, 79, 91]. With highly

Copyright © 2004 by Marcel Dekker, Inc.

112

Chiarizia and Herlinger

aggregated extractants, such as H2MEHP, metal ions are buried within the extractant aggregate and held in place by forces analogous to those which operate in solid ion exchange resins [77]. Similar results have been reported for other highly aggregated extractants, e.g., quaternary alkylammonium salts [92] and dinonylnaphthalene sulfonic acid in nonpolar diluents [93]. All of these effects have been observed to some extent in the metal extraction chemistry of alkylenediphosphonic acid diesters at trace metal concentration level. Several examples are provided in the following with emphasis on the behavior of alkaline earth cations and Am(III). Information on the behavior of other metal ions can be found in the original publications [34, 38, 41, 83, 94–96].

A. Alkaline Earth Cations Solvent extraction data for alkaline earth cations are important for several reasons. Efficient and selective removal of Sr2+ from high level nuclear waste, and the separation of 226Ra and 228Ra from the environment are difficult problems facing the separations community [97, 98]. Also, the coordination chemistry of alkaline earth cations is usually less complicated than transition metal chemistry since the

Figure 5 Acid dependencies for the extraction of selected alkaline earth cations by 0.1 M solutions of H2DEH[MDP], H2DEH[EDP], and H2DEH[BuDP] in o-xylene.

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

113

former is controlled almost exclusively by electrostatic interactions. Therefore, these cations may be used as a convenient tool to probe the solvent extraction chemistry of novel reagents. Figures 5 and 6 show acid and extractant dependencies for the extraction of selected alkaline earth cations by o-xylene solutions of H2DEH[MDP], H2DEH [EDP] and H2DEH[BuDP], respectively In the following discussion it is assumed that the aggregation behavior of the extractants in o-xylene is the same as in toluene, the diluent used for the aggregation studies. The acid dependencies for all extractants exhibit a slope of -2, which is consistent with the displacement of two protons by a divalent cation upon extraction in the organic phase. The extractant dependencies for H2DEH[MDP] and H2DEH [BuDP] are close to 2. Since H2DEH[MDP] is dimeric, each cation is extracted by two H2DEH[MDP] dimers according to the equation: (13)

where H2A represents the fully protonated H2DEH[MDP] extractant. A similar equilibrium can be written for H2DEH[BuDP], with two extractant trimers involved in the extraction equilibrium.

Figure 6 Extractant dependencies for the extraction of selected alkaline earth cations by H2DEH[MDP], H2DEH[EDP], and H2DEH[BuDP] in o-xylene from 0.01 M HNO3 in the aqueous phase.

Copyright © 2004 by Marcel Dekker, Inc.

114

Chiarizia and Herlinger

Structure VIII has been proposed for the organic phase complexes between alkaline earth cations and H2DEH[MDP]:

Structure VIII

In this structure, where for simplicity only one of the monodeprotonated dimers is shown, two protons are displaced from two extractant dimers, as required by the slope analysis. The six-member chelate rings shown in Structure VIII arise from the interaction of the metal ion with the phosphoryl oxygens of the extractant. The extractant dependencies for H2DEH[EDP] have a limiting slope of one in the 0.01–0.1 M concentration range. The slope, however, increases at lower extractant concentrations. As previously shown through Eqs. (9)–(12), a slope of 1 in the log-log plot of metal distribution ratio vs. the analytical concentration of the extractant is consistent with extraction by a highly aggregated species in which the aggregation is not disrupted by the metal ion. This is in agreement with the results from aggregation studies, which indicated that H2DEH[EDP] is primarily hexameric over the concentration range studied. Based on the slope analysis, the extraction of alkaline earth cations by the H2DEH[EDP] hexamer can be expressed as (14)

where (H2A)6 is the H2DEH[EDP] hexamer and MH10A6 is the metal-hexamer complex in the organic phase. In this complex, an alkaline earth cation lies in the hydrophilic cavity formed by the H2DEH[EDP] aggregate, which resembles a reverse micelle. The higher slope values in the 0.001–0.01 M concentration range can be attributed to incomplete aggregation of the extractant. At low extractant concentrations, the measured D values include contributions from extraction by the monomeric extractant. The deviations of the extractant dependency data from a slope one observed in Fig. 6 for the extraction of Ca2+, Sr2-, and Ra2+ were used to calculate the aggregation constant for H2DEH[EDP] reported in Table 5 [94]. A comparison of H2DEH[MDP] and H2DEH[EDP] in Figs. 5 and 6 indicates that the introduction of an additional CH2 group into the alkylene bridge of the diphosphonate profoundly affects alkaline earth extraction. H2DEH[MDP] exhibits no selectivity over the alkaline earth series, while H2DEH[EDP] behaves in a manner

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

115

similar to a monofunctional extractant [79, 99], preferentially extracting Ca. The selectivity exhibited by H2DEH[EDP], however, is accompanied by lower D values than observed with H2DEH[MDP]. The lower D values exhibited by H2DEH[EDP] are consistent with its expected lower acidity and the formation of larger rings upon metal extraction. The silyl-substituted partial esters behave similarly to the dialkyl esters, with the extraction of alkaline earth cations being slightly more efficient. This aspect of the solvent extraction chemistry of the silyl-substituted partial esters will be discussed in more detail for Am(III) extraction.

B. Am(III) Solvent extraction studies with the symmetrical P,P’-disubstituted partial esters were extended to tri-, hexa-, and tetravalent actinides [34, 38, 41, 83, 94, 96]. The behavior of Am(III) is particularly important in view of the need to remove trivalent actinides from high level nuclear wastes. Trivalent actinides are typically left in the waste from the PUREX process which is based on the use of tri-n-butyl phosphate (TBP) as the extractant [100]. Figure 7 shows the acid dependencies for Am(III) extraction by the dialkyland disilyl-substituted esters. Figures 8 and 9 show the extractant dependencies

Figure 7 Acid dependencies for the extraction of Am(III) by (a) 0.1 M solutions of H2DEH[MDP], H2DEH[EDP], and H2DEH[BuDP] and (b) 0.01 M solutions of H2DTMSP[MDP], H2DTMSP[EDP], H2DTMSP[PrDP], H2DTMSP[BuDP], H2DTMSP[PDP], and H2DTMSP[HDP] in o-xylene.

Copyright © 2004 by Marcel Dekker, Inc.

116

Chiarizia and Herlinger

for these solvent extraction systems. Panel C in Figure 9 compares the extractant dependencies for Am(III) extraction measured with the dialkyl and disilyl partial esters of methylenediphosphonic ([MDP]) and ethylenediphosphonic ([EDP]) acids, respectively. The data in Figure 7 indicate that the efficiency of metal ion extraction by the diesters follows the series [MDP]>[EDP]>[HDP]>[BuDP]>[PrDP]>[PDP]. This order does not correspond to the expected trends in basicity, POH acidity, or size of chelating rings possible upon metal ion complexation. This suggests that these factors, as well as the aggregation state of the extractants, have varying degrees of importance and combine in a complex way as the length of the alkylene bridge of the diphosphonate is increased. The acid dependency data for the extraction of Am(III) by H2DEH[MDP] and H2DTMSP[MDP] exhibit a maximum at 0.2–0.3 M HNO3. This feature of the data, which differs from the expected acid dependency for a trivalent metal cation (i.e., a straight line with a slope of -3), probably arises from the formation of species having different stoichiometries at different acid concentrations. In the 0.05–0.3 M HNO3 concentration range, a positively charged species (1:1 ligand to metal complex, for example) can form which requires coextraction of nitrate ions. At higher acidities, a neutral species (2:1 ligand to metal complex) that preferentially reports to the organic phase could be formed, leading to the expected acid dependency slope of -3. The acid dependency data for the disilyl-substituted esters in Fig. 7 exhibit slopes of close to -3 over at least a part of the HNO3 concentration range, consistent with the displacement of three protons upon the extraction of a trivalent metal ion. At higher acid concentrations, however, the extractants with more than three bridging methylene groups tend to exhibit acid dependency plots with a less negative or even positive slope. This suggests an increased importance of extraction by the neutral (fully protonated) extractants via a solvating mechanism, where extraction into the organic phase is dependent on the coextraction of nitrate ions for charge neutralization. This provides additional evidence that the ligands become less acidic as more methylene groups separate the phosphorus atoms. As the bridge length increases, POH acidity decreases and extraction via the solvating mechanism becomes more important at high aqueous acidities. The acidities at which the solvation mechanism manifests itself are generally lower for the ligands with more basic groups. The extractant dependencies for Am(III) extraction by the dialkyl-substituted esters (Fig. 8) show a behavior similar to that discussed previously for the alkaline earth cations. For H2DEH[MDP] and H2DEH[BuDP], the slope values of 2 suggest extraction equilibria of the type: (15) The Am(III) complex with H2DEH[MDP], a ligand for which it can be safely assumed that both phosphonate groups of the molecule cooperatively bind to the same cation (bifunctional behavior), should have a structure similar to that shown

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

117

Figure 8 Extractant dependencies for the extraction of Am(III) by H2DEH[MDP], H2DEH[EDP] and H2DEH[BuDP] in o-xylene from various concentrations of HNO3 in the aqueous phase.

for the alkaline earth cations in Structure VIII, with three protons displaced from two extractant dimers, as required by the slope analysis. The value of unity for the extractant dependency in the extraction of Am(III) by H2DEH[EDP] (Fig. 8), as discussed earlier for the alkaline earth cations, can be explained by the highly aggregated state of the extractant. The extraction equilibrium can be written as (16)

and it has been postulated that Am(III) lies in the hydrophilic cavity formed by the H2DEH[EDP] hexamer. Figure 9 shows the extractant dependencies for Am(III) extraction by the series of disilyl-substituted esters. The plots for the extractants with an odd number of bridging methylene groups (dimers) exhibit extractant dependencies slopes of 2, suggesting that two dimeric aggregates participate in Am(III) extraction. The extraction efficiency decreases as the number of bridging groups increases from 1 to 5, despite the increased basicity of the phosphoryl oxygen along the series. This suggests that the chelate effect may be the dominant factor in determining Am(III) extraction efficiency, with the six-membered rings formed upon complexation by

Copyright © 2004 by Marcel Dekker, Inc.

118

Chiarizia and Herlinger

Figure 9 Extractant dependencies for the extraction of Am(III) by several diphosphonic acid diesters in o-xylene. (a) Data for H2DTMSP[MDP] (1 M HNO3), H2DTMSP[PrDP], and H2DTMSP[PDP] (0.1 M HNO3). (b) Data for H2DTMSP [EDP], H2DTMSP[BuDP], and H2DTMSP[HDP] (0.1 M HNO3). (c) Comparison between H2DTMSP[MDP] and H2DTMSP[EDP] (full symbols) and H2DEH[MDP] and H2DEH[EDP] (empty symbols) (1 M HNO3).

H2DTMSP[MDP] expected to be considerably more stable than the larger rings possible upon Am(III) complexation by H2DTMSP[PrDP] and H2DTMSP[PDP]. However, it should be noted that the acidity of the ligands is expected to decrease over this same series and this may also play a significant role in determining extraction efficiency. The ligands with an even number of bridging methylene groups exhibit Am(III) extractant dependency plots with slopes less than 2, suggesting the importance of extraction by a single, highly aggregated species. H2DTMSP[EDP] exhibits an extractant dependency of 1 and the same considerations made for the H2DEH[EDP] apply. For H2DTMSP[BuDP] and H2DTMSP[HDP], on the other hand, the fractional values for the extractant dependencies arise from simultaneous equilibria involving metal extraction by the monomeric extractant. At the low extractant concentrations used, the concentration of monomeric H2DTMSP[BuDP] and H2DTMSP[HDP] is not negligible and cannot be ignored. Figure 9 (panel C) compares the extractant dependency data for the extraction of Am(III) by H 2 DEH[MDP], H 2 DTMSP[MDP], H 2 DEH[EDP], and

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

119

H2DTMSP[EDP]. While the disilyl esters exhibit nearly identical aggregation behavior as their dialkyl analogues, the disilyl esters extract metal ions two to three times more efficiently. As described earlier, the increased extraction efficiency is due to the slightly higher basicity in the disilyl esters due to the increased electron donating nature of the TMS group. Because of the interplay between extractant structure and diluent properties, the extraction behavior of diphosphonic acid diesters is expected to change significantly if a more polar diluent such as 1-decanol is used. Due to hydrogen bonding between these extractants and the diluent, only monomeric solute species are observed in 1decanol [66]. An investigation of the extraction of alkaline earth and actinide cations from aqueous nitric acid into 1-decanol solutions of H2DEH[MDP], H2DEH[EDP], and H2DEH[BuDP] was initiated to compare the solvent extraction chemistry of these reagents in the absence of aggregation effects [95, 96]. The extractant dependencies observed for Am(III) and the other cations with H2DEH[EDP] in 1-decanol are not unity. In this diluent, metal ions bind to a number of monomeric extractant molecules depending upon cationic charge (up to three with Am(III)), analogous to monomeric monofunctional extractants [91]. This confirms that the extractant dependencies of unity observed for the extraction of metal ions by H2DEH[EDP] in nonpolar diluents arise from the highly aggregated state of the extractant. Extraction of Am(III) by H2DEH[MDP], H2DEH[EDP], and H2DEH [BuDP], in the aqueous acidity region where an acid dependency of -3 applies, can be described by the simultaneous equilibria (17) and (18), which are analogous to those describing the extraction of actinide ions by HDEHP [101]: (17) (18) Equilibrium (18) is stoichiometrically indistinguishable from equilibrium (19): (19) Equilibria (18) and (19) imply different structures for the metal species in the organic phase [96]. Similar equilibria can be written for the extraction of alkaline earth cations by the three ethylhexyl-substituted extractants in 1-decanol [95]. The most striking feature of the data is the extent to which metal ion extraction is suppressed by changing the diluent from 0-xylene to 1-decanol. In most cases, metal extraction is reduced by several orders of magnitude [96]. Extraction of metal ions by 1-decanol solutions of the ligands is suppressed relative to that in oxylene because the alcohol is strongly hydrogen-bonded to the phosphoryl groups of the acids. This makes the groups of the ligand unavailable for extractant aggregation and provides competition with the metal ion for ligand binding sites.

Copyright © 2004 by Marcel Dekker, Inc.

120

Chiarizia and Herlinger

The reduced extraction efficiency in 1-decanol can provide a facile route for actinide stripping. Recovery of actinide ions from an o-xylene phase containing H2DEH[MDP] is very difficult because of the extremely strong affinity of the extractant for these metal ions. Stripping could be accomplished much more readily by diluting the loaded organic phase with 1-decanol or a similar alcohol. Although the extraction properties of diphosphonic acid esters in 1-decanol/o-xylene mixtures have not been investigated, it is reasonable to assume that conditions for substantially reducing actinides distribution ratios at conveniently low aqueous acidities could be found, thus permitting facile stripping based solely on diluent effects.

VI. SOLVENT EXTRACTION OF METAL IONS AT HIGH LOADING The solvent extraction data discussed in Section V were obtained at tracer metal concentration levels. Although distribution measurements at very low metal concentrations are essential for establishing the stoichiometry of metal extraction through graphical slope analysis and other methods, the information provided by these studies generally cannot be extrapolated to pratical solvent extraction conditions where the metal concentrations are much higher. Very little information exists in the literature on the organic phase speciation at high metal loading. When the concentrations of extracted metal approach those corresponding to saturation of the extractant in the organic phase, the discrete complexes familiar in solution coordination chemistry tend to disappear in favor of self-assembled structures. The driving force for the aggregation of the electroneutral metalextractant complexes is generally provided by van der Waals attraction between polar solutes in low-polarity diluents. As mentioned earlier, the technique of smallangle neutron scattering (SANS) is well suited for studies on the morphology of these aggregates. For example, self-assembly of the metal-extractant complexes in large cylindrical structures having lengths of hundreds of Å’s have been observed through SANS measurements for HDEHP and other monofunctional organophosphorus acidic extractants after extraction of Co(II) at high concentrations [87]. Besides being a fascinating problem in physical and structural chemistry, such species are of considerable technological importance. Recent studies have correlated the formation of aggregates in organic solutions of extractants with the formation of a third organic phase, a phenomenon still largely unexplained from a structural standpoint [102, 103]. Bifunctional extractants such as diphosphonic acid partial esters are expected to exhibit a strong tendency to aggregate or polymerize under the influence of high metal concentrations, as metal ions can promote the formation of large polynuclear species by bridging functional groups of different extractant molecules. This behavior was suggested by some peculiarities of the solvent extraction data at

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

121

low metal loading, and confirmed by VPO measurements on H2DEH[MDP] solutions in toluene containing high concentrations of various metals. Numberaverage aggregation numbers as high as 10, 14, and 33 were obtained for U(VI), Th(IV), and Fe(III) solutions, respectively [65]. To gain more information on the morphology of the species present in organic phases containing diphosphonic acid diesters after extraction of a variety of metal ions at high concentrations, SANS measurements were performed using deuterated toluene as the diluent [80, 81, 88]. Representative results are reported in Table 6 for H2DEH[MDP], H2DEH [EDP] and H2DEH[BuDP]. For each system investigated, the table reports the HNO3 concentration in the aqueous phase and the concentrations of diester and metal in the organic phase. Typically, these metal concentrations are the highest attainable under the conditions used for metal extraction. The values of the diester to metal concentration ratios in the organic phase give an indication of how close each system is to extractant saturation. The radius of gyration Rg, provided by the Table 6 Radius of Gyration Rg and Weight-Average Aggregation Number nw for Solutions of H2DEH[MDP], H2DEH[EDP] and H2DEH[BuDP] After Extraction of Metal Ions at High Concentrations

a b

Maximum value Solution obtained by dissolving 33.0 mg Fe2(DEH[MDP])3 in 1 mL deuterotoluene.

Copyright © 2004 by Marcel Dekker, Inc.

122

Chiarizia and Herlinger

Guinier analysis of the data, is a measure of the spatial extension of the aggregate and is given by the root-mean-squared distance of all the atoms from the center of gravity of the scattering particle. The weight-average aggregation number of the extractant nW, calculated from the SANS data, gives the number of diester molecules contained in each aggregate.

A. Metal Species Formed by H2DEH[MDP] The results in Table 6 show that very large aggregates are formed in the extraction of Th(IV) and Fe(III) at high concentrations. The Th(IV)–H2DEH[MDP] system is polydisperse with a number of aggregates of different size contributing to the overall scattering. The largest particles that can be observed under the experimental limitations of the SANS measurements [88] exhibit Rg and nw values of 87 Å and 190. The modified Guinier analysis for rod-shaped particles was performed on the Th(IV)–H2DEH[MDP] data. This analysis allows an estimation of the radius of the rodlike structure from the slope of the straight line describing the SANS data in the form {ln[I(Q)·Q] vs. Q2} [104]. The modified Guinier analysis revealed the existence in solution of cylindrical aggregates about 300 Å long with a radius of about 10 Å. The analysis of the data, however, also revealed the existence of larger aggregates which grow simultaneously in all directions. It is likely that crosslinking occurs to form particles analogous to Structure IX (where the double bonds have been omitted for simplicity):

Structure IX

A structure similar to Structure IX was hypothesized previously for the polymers formed by tetravalent uranium with dialkylpyrophosphates [105], ligands that are structurally similar to H 2DEH[MDP], but have the two phosphorus atoms bridged by an oxygen atom instead of a methylene group. A

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

123

remarkable feature of the Th(IV)–H2DEH[MDP] system is that large aggregates form only at the highest possible metal concentration in the organic phase. Attempts to further increase the concentration of Th(IV) result in the precipitation of the Th(DEH[MDP])2 salt. Structure IX probably represents the formation in solution of large structural units having the same threedimensional structure as the Th(DEH[MDP])2 salt. The Fe(III)–H2DEH[MDP] system was investigated in great detail [80]. Table 6 summarizes the results obtained for the solutions closest to extractant saturation. Figure 10 shows the SANS data in the form of Guinier plots for the samples obtained by extracting Fe(III) from 0.1 M HNO3. The progressive increase of scattering for increasing Fe(III) concentrations is clearly visible. Figure 11 shows the nw aggregation number of the aggregates as a function of the organic phase Fe(III) concentration. As more Fe(III) is transferred into the organic phase and the extractant saturation is approached, the nw values grow dramatically independently of the aqueous phase acidity. In all cases, the scattering particles for solutions close to extractant saturation contain from about 70 to about 110 H2DEH[MDP] molecules. Modified Guinier analysis for rod-shaped particles demonstrated that these aggregates are cylindrical with lengths up to about 300 A. However, unlike the largest Th(IV)–H2DEH[MDP] aggregates, the radius of the Fe(III)–H2DEH[MDP]

Figure 10 Guinier plots of the SANS data for 0.1 M H2DEH[MDP] in deuterated toluene after extraction of Fe(III) from 0.1 M HNO3. (From Ref. 80.)

Copyright © 2004 by Marcel Dekker, Inc.

124

Chiarizia and Herlinger

Figure 11 Weight-average aggregation number, nw, and ratio of H2DEH[MDP] to Fe(III) concentration in the organic phase as a function of organic Fe(III) concentration and aqueous phase acidity. (From Ref. 80.)

aggre gates is always constant at about 9 Å, suggesting that particle growth takes place only along the long axis of the cylinder. To obtain more information on these “wormlike” aggregates, SANS measurements were performed on a sample prepared by direct dissolution of the Fe2(DEH[MDP])3 salt in deuterated toluene. The SANS data revealed the presence of cylindrical particles which are similar to those observed for samples prepared through Fe(III) extraction. In these particles, the hydrocarbon chains of the diester should be oriented toward the exterior of the cylindrical aggregates, while the metal ions interact with the polar groups of the extractant which are oriented toward the interior of the cylinder. The metal ions thus are located along a channel in the center of the cylinder. In this case, a likely structure of the metal-extractant polymeric species is shown in Structure X. This structure is consistent with the 3:2 stoichiometric ligand to metal ratio in the salt, and the preferred octahedral coordination environment of Fe(III). Similar structures have been reported for solid-state, eight- and seven-coordinate, crystalline lanthanide complexes of 1-hydroxyethane-1,1-diphosphonic acid (HEDPA), an aqueous soluble, unsubstituted analog of H2DEH[MDP] [106]. From the ligand to

Copyright © 2004 by Marcel Dekker, Inc.

Symmetrical P,P’-Disubstituted Esters

125

metal stoichiometric ratio and the information provided by the Guinier analysis, the Fe-Fe distance in Structure X was calculated to be 3.7±0.3 Å.

Structure X

Structure X and a Fe-Fe distance of 3.7 Å suggest formation of covalent bonds between Fe(III) and the ligand molecules. Covalent binding of Fe(III) to H2DEH[MDP] was independently demonstrated in an investigation of the farinfrared spectrum of Fe2(DEH[MDP])3 [69]. Based on the information provided by IR and SANS measurements, it seems reasonable to conclude that, at least for the Fe(III)-H2DEH[MDP] system, the aggregation is more likely to be an actual polymerization process than the formation of aggregates where the individual molecular units are held together by a purely physical interaction.

B. Metal Species Formed by H2DEH[EDP] and H2DEH[BuDP] The nw values measured for H2DEH[EDP] solutions after extraction of Ca2+, La3+, and are not significantly different from that of the extractant alone. These results confirm that metal extraction occurs through cation transfer into the hydrophilic cavity of a hexameric aggregate with little, if any, disruption of the solution structure of the extractant. Infrared spectra indicate that extraction of Th(IV) and Fe(III) profoundly alter the solution structure of H2DEH[EDP] [37]. The nav values in Table 6 are consistent with the hypothesis that high concentrations of Th(IV) and Fe(III) in the organic phase disrupt the hexameric extractant aggregation and result in the formation of different metal extractant species. Formation of large aggregates was observed for H2DEH[BuDP] only upon Fe(III) extraction [81]. When the organic phase is fully loaded with Fe(III), these aggregates contain up to 85 H2DEH[BuDP] molecules. The modified Guinier analysis for rod-shaped particles confirmed that the Fe(III)-H2DEH[BuDP] aggregates are cylindrical. However, a fundamental difference exists between the aggregates formed by Fe(III) with H 2DEH[BuDP] and those formed with H2DEH[MDP]. The Fe(III)H2DEH[MDP] aggregates described earlier have a constant radius and grow length-wise with increasing Fe(III) concentration. In contrast, the Fe(III)-H2DEH[BuDP] aggregates have an approximately constant

Copyright © 2004 by Marcel Dekker, Inc.

126

Chiarizia and Herlinger

length (~60 Å), but a radius that increases significantly with the concentration of Fe(III) in the organic phase (up to ~15 Å). We conclude this section by emphasizing that also from the standpoint of self-assembly of metal-extractant complexes in the organic phase, the diphosphonic acid diesters exhibit a variety of behaviors that depend on the charge and coordination chemistry of the metal ion, as well as on the aggregation behavior of the diester. For example, the behavior of heavily metal loaded H2DEH[EDP] solutions is dominated by the presence of spherical hexameric extractant aggregates that in many cases retain their morphology even in the presence of high metal concentrations. The behavior of H2DEH[MDP] and H2DEH[BuDP] solutions close to the point of saturation, on the other hand, is characterized by the tendency to form long cylindrical polymeric species (especially with Fe(III) and Th(IV)), whose thickness depends on the specific metal ion-extractant system. While it is difficult to explain the differences in the behavior of Fe(III) with the three extractants, especially the absence of large aggregates with H2DEH[EDP], it is remarkable that small changes in the distance between the phosphorus atoms of the ligand can lead to such profound differences in the aggregation of the extracted complexes.

VII. ENTHALPY AND ENTROPY CHANGES IN METAL SOLVENT EXTRACTION When a cation is transferred from an aqueous medium into an organic phase through an ion exchange complexation mechanism, the net enthalpy and entropy changes associated with the extraction mainly result from four opposing processes: (1) the dehydration of the extracted cation and the hydration of the exchanged protons; (2) the metal coordination by the organic ligand and the deprotonation of the ligand [107]. For ligands aggregated through hydrogen bonds, as typical for organophosphorus acids, this second process also involves breaking hydrogen bonds. Dehydration of the cation generally involves a positive enthalpy change (∆H> 0) as a result of breaking ion-water bonds, and a positive entropy change (∆S>0) due to the increased disorder of the system. The opposite changes will occur for the hydration of the proton. Metal coordination by the organic ligand will result in a negative enthalpy change (∆HH2DEH [BuDP]. Although the thermodynamic parameters measured for H2DEH[1,2-BzDP] cannot be directly compared with those for the other partial esters due to the presence of significant amounts of 2-ethylhexanol in solution, the superior efficiency of H2DEH[1,2-BzDP] compared to H2DEH[MDP] reflects the effect of preorganized ligand binding groups. This superiority is somewhat surprising since a comparison with carboxylate ligands of the same geometry shows that malonic acid is a stronger complexant of Eu3+ than phthalic acid [115]. The ∆H° values for Am3+ extraction increase along the H2DEH[MDP]< H2DEH[EDP]2-ethylhexyl p-tolyl sulfoxide (1.46)>di(2-octyl) sulfoxide (1.12)>di(p-ethylphenyl) sulfoxide (0.24), being in accord with the trend revealed by the preceding sequence. However, at 4 M HNO3 the sequence is 2-ethylhexyl p-tolyl sulfoxide (4.41)>DEHSO (3.16)>di(pethylphenyl) sulfoxide (1.21)>di(2-octyl) sulfoxide (0.20) [11]. Cyclooctyl octyl sulfoxide [33] and cyclohexyl butyl sulfoxide [32] are in xylene less effective than dialkyl sulfoxides, while cyclohexyl hexyl sulfoxide and dicyclohexyl sulfoxide are more effective [15]. Also in xylene, 2-heptylthiophane S-oxide [30, 32, 15], 3-hexylthiophane S-oxide [32], 2-butyl-5-methylthiophane S-oxide [57], 2-propyl-4-isopropylthiophane S-oxide [32], 2(cyclohexylmethyl)thiophane S-oxide [30], 2-butylthiacyclohexane S-oxide [57], 2-pentylthiacyclohexane S-oxide [15], and 2-methyl-1-thiadecalin S-oxide [30] are more effective than dialkyl sulfoxides, while 2-thiadecalin S-oxide [30] and hexyl phenyl sulfoxide [15] are less effective. Petroleum sulfoxides are mixtures of thiaheterocyclic S-oxides and their composition, together with their extracting power, is dependent on their origin. For example, in kerosene, they may be more effective than dialkyl sulfoxides, as indicated in Fig. 2, or they may be comparably or less effective [43, 44, 46, 47]. In the pair DHxSO/2-heptylthiophane S-oxide the higher effectiveness of the latter is ascribed to the entropy factor [32]. The extraction ability of both cyclic and noncyclic sulfoxides is well correlated with their basicity, and that of noncyclic compounds is well correlated with the sum of the electronegativities of the substituent groups [15]. Surprising is the weak extraction ability of bifunctional extractants. One of them is bis(octylsulfinyl)methane, which at a concentration of 0.5 M in 1,1,2,2tetrachloroethane gives DU(VI) values of >1 only at >4 M HNO3. A maximum DU(VI) value of merely 5.5 is attained at 8.5 M HNO3 [58]. The other is phenyl-N,Ndibutylcarbamoylmethyl sulfoxide. In the extraction of initially 0.005 M U(VI) by a 0.3 M solution of the compound in toluene, the DU(VI) value increases from 0.09 at 1 M HNO3 to 1.4 at 6 M HNO3 [59]. Comparison with curve 4 in Fig. 2 shows that the compounds extracts U(VI) much less effectively than DEHSO. More effective is a trifunctional extractant, namely, bis(N,N-dioctylmethylcarbamoyl) sulfoxide. A 0.2 M solution in dodecane yields in the extraction of trace U(VI) the values DU(VI)=14 and 1.4 at 1 and 9.5 M HNO3, respectively [60]. Rough comparison with curve 3 in Fig. 2 indicates that the trifunctional compound is a more powerful extractant for U(VI) than DEHSO. In comparison with other extractant types, the ability to extract uranyl nitrate into benzene or CCl4 as a disolvate at a low concentration of HNO3 increases in the series TBPDOSO~ cyclooctyl octyl sulfoxide, all in xylene and with 0.02 M HNO3+4 M NaNO3 as the aqueous phase [33]. The effect of the sulfoxide structure can be dependent on the HNO3 concentration, as shown by the extractant (DTh) sequences DEHSO (0.67) >di(p-ethylphenyl) sulfoxide (0.44)>di(2-octyl) sulfoxide (0.12)>2-ethylhexyl ptolyl sulfoxide (0.094) at 1 M HNO3 and di(2-octyl) sulfoxide (19.6)>DEHSO (1.31)>2-ethylhexyl p-tolyl sulfoxide (0.59)>di(p-ethylphenyl) sulfoxide (0.1) at 4 M HNO3 (al in xylene) [11]. 2-Heptylthiophane S-oxide is more efficient than dialkyl sulfoxides [75]. Pu(IV) is very effectively extracted by bis(N,N-dioctylcarbamoylmethyl) sulfoxide. A 0.2 M solution of the compound in dodecane yields at 1.5-9 M HNO3 DPu(IV) values as high as 100–220, with a maximum attained at 4 M HNO3. So Pu(IV) is extracted selectively over U(VI), the separation factor ␣Pu(IV)/U(VI) being 7, 19, and 63 at 1, 5, and 9 M HNO3, respectively [60]. A dependence of the structure effect on the acid concentration was found also in the extraction of Zr(IV). DEHSO was more efficient than di(2-octyl) sulfoxide at 1 M HNO3, but it is less efficient at 4 M HNO3 [11]. To compare with other extractant types, the ability to extract Th(IV) nitrate at a low HNO3 concentration into benzene as a trisolvate increases in the series TBP< DOSOCCl4 (3.4)>benzene (2.2) [81]. The extraction of Pu(IV) is strongly suppressed by alcohol modifiers which may be used to prevent the formation of a third liquid phase. If 0.4 M DEHSO in dodecane contains 3, 5, and 20 vol% 2-ethylhexanol, the DPu(IV) value in the extraction from 2 M HNO3 is lowered from 9.0 without modifier to 4.1, 2.1, and 0.61, respectively. With isodecanol the DPu(IV) value is decreased to 3.8, 2.0, and 0.85, respectively [62]. To be noticed is the effect of water-miscible organic additives to the aqueous phase. As observed in the extraction of trace Pu(IV) by 0.2 M DEHSO in dodecane, the effect of organic additives is the highest at 20 vol%, and such an addition to the aqueous phase essentially changes the form of the DPu(IV) vs. [HNO3] dependence. The round maximum at 5 M HNO3 is changed to a very sharp one at 2.1 M HNO3, and the DPu(IV) value at this acidity is increased by a factor of 3.9 by acetonitrile, 3.2 by acetone, 2.6 by ethanol, 2.3 by methanol, and 1.8 by dioxane. Propanol suppresses the DPu(IV) value [78]. An unlike effect was found in the extraction by DBSO and DiPSO in Solvesso 100 at 20–50% additive. Also at 2 M HNO3, with 0.2 M DBSO the DPu(IV) value is changed by a factor of 1.4–2.2 by acetone, 1.6–2.9 by acetonitrile, 0.73–0.59 by dioxane, 1.0–0.57 by methanol, 1.3–0.52 by ethanol,

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

189

and 0.43–0.21 by 1-propanol. With 0.2 M DiPSO the factors are 3.2–4.2 by acetone, 3.2–3.8 by acetonitrile, 1.7–1.1 by dioxane, 2.5–0.76 by methanol, 2.4–1.2 by ethanol, and 0.82–0.37 by 1-propanol [29]. 3. Nature of Extracted Complexes There is evidence by slope analysis or indication by saturation experiments for the formation of disolvates of the type M(NO3)4 · 2B, with few exceptions for Th(IV) and Zr(IV). In the case of Zr(IV) the disolvates could also possibly be Zr(NO3)2(OH)2 · 2B or Zr(NO3)4 · 2HNO3 · 2B. The disolvates have been found in the extraction of • • • • •

Th(IV) by DBSO in xylene from 2 M HNO3 [74], DPSO in TCE [22] and DOSO in benzene in the absence of HNO3 [76], DEHSO in kerosene from 3.5 M HNO3 [11, 28], and PetrSO in kerosene from 3 M HNO3 [46, 47] Np(IV) by DiPSO, DHxSO, and DOSO in Solvesso 100 from 2 M HNO3 [25] Pu(IV) by DHxSO [26] and DOSO [38] in Solvesso 100, and DEHSO in dodecane [78], all from 2 M HNO3 Zr(IV) by DHxSO in xylene from 4 M HNO3 [75] and in Solvesso 100 from 2 M HNO3 [26], and DOSO in CCl4 from 3.7–7.9 M HNO3 [80] Hf(IV) by DBSO in cyclohexane from 8 M HNO3 [81]

Contrary to these reports, • • •

A trisolvate of the type Th(NO3)4 · 3B was found to be extracted by DOSO, DEHSO, di(2-octyl) sulfoxide, and cyclooctyl octyl sulfoxide from 0.02 M HNO3+4 or 6 M NaNO3 [33]. A trisolvate of the type Hf(NO3)4 · 3B was found to be extracted by DPSO in CCl4 from 9 M HNO3 [82]. A monosolvate of the alleged composition ZrO(NO3)2 · B was found in the extraction of Zr(IV) by DBSO from 7 M HNO3 [79].

A slight effect of the Zr(IV) concentration on DZr values was observed in the extraction of Zr(IV) by 40% DBSO in xylene from 7 M HNO3. The DZr value increased from 2.8 at 5.5×10-4 M Zr(IV) merely to 6.7 at 0.055 M Zr(IV) [79]. The experimental points scattered considerably and the increase was not an unambiguous proof of partial self-association of the extracted Zr(IV) complex. The DHf values decrease with increasing Hf(IV) concentration in the extraction by 1 M DHpSO in TCE from 7–10 M HNO3. At 10 M HNO3 the DHf value is lowered from ~1 to ~0.1 when the initial aqueous Hf(IV) concentration is increased from 5.6×10-4 to 0.017 M [34]. Thus, the suppression cannot be ascribed to a decrease of the free extractant concentration. The phenomenon was not commented on in the original source and, without a more detailed study, it does not suffice for evidencing dimerization of Hf(IV) in the aqueous phase.

Copyright © 2004 by Marcel Dekker, Inc.

190

Kolarik

The Zr4+ ion is bound to the O atom of the SO group: The respective band of DOSO in CCl4 is shifted by Zr(IV) from 1032 to 934 cm-1 [80]. 4. Thermodynamic Functions Published values are gathered in Table 6. In Refs. 73, 74 they were determined from the dependence of the concentration equilibrium constant of the extraction on 1/T, in other work from the temperature dependence of the distribution ratio of M(IV). The nitrate complexing of M(IV) in the aqueous phase and the extraction of HNO3 together with M(IV) were taken into account in Refs. 72, 73.

D. Miscellaneous Metals 1. Distribution Data Selected sources of distribution data are given in Table 7. It is seen that little attention has been paid to the extraction of trivalent, pentavalent, and bivalent elements. Nb(V) and Ru, the latter in an unspecified valency state, are weakly extracted by 0.2 M DEHSO in kerosene from 0.7 to 4 M HNO3. The maximum values are DNb(V)~0.004 at ~3 M HNO3 and DRu(?)~0.012 at ~1.5 M HNO3 [11]. The good extractability of Pd(II) by DEHSO and the weak extractability of a series of metals by DBSO is shown in Fig. 7 where the extraction of Th(IV) is shown for comparison. Fe(II) and Al(III) are slightly less extractable than Pb(II), the extractability of Zn(II) and Mn(II) is similar to that of Ni(II), the extractability

Table 6 Thermodynamic Values of the Extraction of M(IV) in the Form of the Complex Th(NO3)4·2B

a

With 4 M NH4NO3 also present.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

191

Table 7 Survey of Data on the Distribution of Miscellaneous Metals Between Solutions of Sulfoxides and Aqueous Nitrate Solutions

of Mg(II) and Ca(II) is similar to that of Sr(II), and Na(I) and K(I) are still less extractable than Sr(II). Fe(III) is moderately extracted by 1 M DHpSO in TCE, with DFe(III) increasing from 0.017 at 2 M HNO3 to 1.0 at 8 M HNO3 [34]. 2. Effect of Extraction Variables Am(III) and lanthanides(III) are so weakly extractable by monofunctional sulfoxides that relevant extraction efficiency is attained only at high concentrations of salting out agents. These must be salts of nonextractable metals, such as alkali or alkaline earth nitrates. Nitric acid cannot be used as a salting-out agent, because it strongly suppresses the extraction. For example, in the extraction of Am(III) by 0.4 M

Copyright © 2004 by Marcel Dekker, Inc.

192

Kolarik

Figure 7 Extraction of various metals (initially di(2-octyl) sulfoxide>di(3-octyl) sulfoxide [16]. A very similar picture was obtained in the extraction by 0.5 M DOSO, DEHSO, and di(2-octyl) sulfoxide in xylene from 0.28 M HNO3+6 M LiNO3 [7]. By the way, data in Fig. 8 show the tetrad effect to a limited extent only, and data given in Ref. 7 do not do it much more visibly.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

193

Figure 8 Extraction of Y(III) and lanthanides(III) by 0.4 M isomeric dioctyl sulfoxides in xylene. Curve 1—DOSO; 2—DEHSO; 3—di(2-octyl) sulfoxide; 4—di(3-octyl) sulfoxide. Initial aqueous phase: all elements (each 0.005 M), 5.0 M LiNO3, pH 2.0. 20°C. (From Ref. 16.)

Bifunctional bis(octylsulfinyl)methanes and 1,2-bis(octylsulfinyl)ethanes with various octyl isomers are much more effective extractants for lanthanides(III) than the corresponding dioctyl sulfoxides. The ethane derivatives are more efficient than the methane derivatives, and within each extractant group the extraction ability for Eu(III) decreases in the octyl order n-octyl>2-ethylhexyl>2-octyl (0.2 M extractants in chloroform, initially 0.06 M Eu(III)+6 M NH4NO3+0.3 M HNO3). The bi-functional compounds give lower separation within the lanthanide(III) group than dioctyl sulfoxides [7]. High extraction efficiency for Am(III) is exhibited even in the absence of a salting out agent by the trifunctional extractant bis(N,N-dioctylcarbamoylmethyl) sulfoxide. The DAm(III) value increases from 6.3 at 3 M HNO3 to 28 at 9 M HNO3 (0.2 M extractant in dodecane) [60]. The position of sulfoxides among diverse extractants for Tm(III) is exemplified by two sets of data. They imply that not only the basicity of the oxygen donor atom but also the diluent and the ionic medium in the aqueous phase can influence the order of increasing extraction ability. The extraction of Tm(III) nitrate into CCl4 from 5.9 M Al(NO3)3 at pH 1.0 increases in the order trioctylamine N-oxidedodecane (12.0)>Solvesso 100 (10.0)>CCl 4 (9.2)> cyclohexane (8.0)>benzene (7.4) ~o-dichlorobenzene (7.3)>1,2-dichloroethane (5.0)>xylene (3.74)>chloroform (1.96) [85]. The extraction of Am(III) is enhanced by adding a water-miscible organic additive. In the extraction by DEHSO in dodecane and at additive concentrations of 30–40 vol%, the extent of the enhancement decreases in the additive order acetonitrile >acetone>methanol>dioxane>ethanol. Propanol suppresses slightly the extraction [83]. Similar results were obtained with DiPSO and DOSO [86]. 3. Nature of the Extracted Complexes Trisolvates of the type M(NO3)3·3B were found in the extraction of • • • •

Y(III), La(III), Gd(III), and Lu(III) by DHxSO, ethyl octyl sulfoxide, DOSO, DEHSO, di(2-octyl) sulfoxide, di(3-octyl) sulfoxide and cyclooctyl octyl sulfoxide in xylene from 5 M LiNO3 at pH 2 [16] Nd(III) and Tm(III) by DOSO in CCl4 from 6 M LiNO3 at pH 3.0 [84] La(III) by DOSO and DEHP in probably xylene from an unspecified acidic LiNO3 solution [7] Am(III) by DEHSO in dodecane from 0.03 M HNO3+3 M Ca(NO3)2 [83]

Disolvates of the type M(NO3)3·2B or M(NO3)2·2B were found in the extraction of • • •

La(III) by 1,2-bis(2-octylsulfinyl)ethane in xylene from an unspecified solution [7] Pd(II) by DEHSO in dodecane, benzene, and toluene from 2 M HNO3 [85] Co(II) by DOSO in benzene from Co(NO3)2 [87]

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

195

Some fraction of M(NO3)3·4B may be present in the extraction of La(III) by DHxSO, ethyl octyl sulfoxide, DEHSO, di(2-octyl) sulfoxide, and di(3-octyl) sulfoxide, of Gd(III) by di(2-octyl) sulfoxide, and of Y(III) and Lu(III) by di(2octyl) sulfoxide and di(3-octyl) sulfoxide, all in xylene. A nitrate to M(III) ratio has been determined as 3.0±0.2 in the extraction of Nd(III), Er(III), and of Y(III) by DHxSO in xylene [16]. A tetrasolvate of the type Co(NO3)2·4B is extracted by DOSO in heptane, CCl4, and chloroform [87]. The water content of the organic phase loaded with lanthanides(III) was reported in Ref. 7 but, unfortunately, the composition of the phase was not given and no picture of the water to lanthanide(III) or water to extractant ratios can be obtained. The average water content decreases in the order bis(2-octylsulfinyl) methane>DOSO>DEHSO>di(2-octyl) sulfoxide. An intensive water band was also found in the infrared spectrum of the solid complex of 1,2-bis(2-ethylhexylsulfinyl) ethane with Pr(NO3)3, and it was suggested that the band belonged to coordinated water [7]. Ir spectra show that the trisolvate Nd(NO3)3·3B with B=DOSO is anhydrous in CCl4. A shift of the band of the SO group from 1050 to 1002 cm-1 is evidence that the Nd3+ ion is bonded to the O atom of the SO group [84]. Contrarily, the Pd2+ ion is bound to the S atom of DEHSO, as substantiated by a 190 cm-1 shift of the S=O stretch toward higher frequency [85].

III. EXTRACTION FROM CHLORIDE MEDIA A. Hydrochloric Acid HCl is generally weakly extractable by symmetrical sulfoxides. For example, the concentration of HCl in 0.1 M DHxSO in benzene is 8×10-4 and 0.062 M after contact with 4 and 10 M HCl, respectively [20]. A HCl concentration of ~3×10-4 M is attained in 0.02 M ethyl dodecyl sulfoxide in xylene after contact with 6 M HCl. The extraction of HCl is visibly weaker than that of HClO4 and much weaker than that of HNO3 [3]. 0.1 M methyl 4,8-dimethylnonyl sulfoxide in p-xylene extracts HCl rather strongly. An acid to sulfoxide ratio of 1.7 is attained in the organic phase at 9 M HCl. Formation of a sesquisolvate or of a mixture of mono- and disolvate is indicated at 1–8 M HCl, but the picture may be obscured by a high solubility of the extractant in the aqueous phase (the DB value is as low as 4 at 8 M HCl) [8].

B. Hexavalent and Pentavalent Metals 1. Distribution Data U(VI) and Pa(V) belong to often studied metals, even if chloride solutions are not compatible with the anticipated use of sulfoxides in nuclear processes where a

Copyright © 2004 by Marcel Dekker, Inc.

196

Kolarik

nitrate medium is preferred. A survey of important sources of distribution data is given in Table 8. 2. Effect of Extraction Variables Hydrochloric acid is the most frequently used salting-out agent. Figures 9 and 10 illustrate different forms of the DM vs. [HCl] dependencies, showing that many of them exhibit a maximum at different HCl concentrations. In the extraction of U(VI)

Table 8 Survey of Data on the Distribution of Hexa- and Pentavalent Metals Between Solutions of Sulfoxides and Aqueous Chloride Solutions

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

197

Figure 9 Effect of HCl on the extraction of trace Pa(V) and U(VI) by various sulfoxides. Curve 1—Pa(V) by 0.25 M DOSO in CCl4; 2—Pa(V) by 0.25 M DPSO in CCl4; 3—U(VI) by 0.25 M DPSO in CCl4; 4—U(VI) by 0.25 M DOSO in CCl4; 5—U(VI) by 0.25 MDDSO in CCl4. Room temperature. (From Ref. 88.)

Figure 10 Effect of HCl on the extraction of Cr(VI), Mo(VI), and W(VI) by various sulfoxides. Curve 1—initially 0.001 M Cr(VI) by 0.1 M DBSO in benzene [50]; 2—trace Mo(VI) by 10% PetrSO in toluene [90]; 3—initially 0.0005 M Mo(VI) by 0.2 M DOSO in benzene [76]; 4—initially 0.0008 M W(VI) by 0.2 M DOSO in benzene [76]; 5—trace W(VI) by 10% PetrSO in toluene [90]. Room temperature.

Copyright © 2004 by Marcel Dekker, Inc.

198

Kolarik

by DPSO to DDSO the maximum occurs at 7–9 M HCl. It is slightly shifted toward higher HCl concentrations with increasing alkyl length [88]. The distribution ratios increase continuously with the HCl concentration (0.5–10 M) in the extraction of Nb(V) and Ta(V) by 0.5 M DOSO in benzene. Nb(V) is extracted preferably to Ta(V) at 3–10 M HCl [76]. To view the effect of the extractant structure, DU(VI) and Dpa(v) values were measured with DHxSO to DDSO in CCl4 at 7 M HCl and plotted vs. the molar mass of the five sulfoxides. They lie on a curve with a maximum at DOSO [93]. However, more extensive data on U(VI) are given in Ref. 88 and they show that this is a simplified picture. In fact, the DU(VI) vs. [HCl] dependences for DPSO to DDSO do not reveal any simple trend in the influence of the sulfoxide alkyl chain length. Three of the curves are shown in Fig. 9. The only regularity is that the position of the maximum is shifted to higher HCl concentrations at increasing chain length. However, the height of the maximum decreases in the order DOSO>DPSO>DNSO>DHpSO>DDSO>DHxSO. Also the width of the maxima changes with the chain length and, thus, the dependencies cross to the left and right of the maxima and different relations between the extraction efficiency and the chain length are found at different HCl concentrations. An obvious order can be seen only at 4-6 M HCl, namely DPSO>DOSO>DHxSO ~DHpSO~DNSO~ DDSO. TBP extracts U(VI) from HCl less efficiently than DPSO and DOSO in CCl4 [94], and the order DOSO>DPSO>TBP (all in CCl4) has been reported for Pa(V) [88]. Also, Mo(VI), W(VI), Nb(V), and Ta(V) are extracted by TBP more weakly than by DOSO, both in benzene [76]. The diluent effect has been investigated in the extraction of U(VI) by 0.25 M DPSO from 3 M HCl. The DU(VI) value decreases in the diluent (DU(VI)) order benzene (1.23)>toluene (0.85)>xylene (0.67)>CCl 4 (0.63)>butyl acetate (0.28)> cyclohexanol (0.19)>1-pentanol=benzyl alcohol (0.15). A linear relationship between the logarithms of DU(VI) and the dielectric constant of the diluent is fairly well followed [94]. The same relationship has been found in the extraction of Pa(V) by DPSO from probably 3 M HCl, but for some of the studied diluents only. Here the diluent order is CCl4>CS2>chloroform>chlorobenzene>1-pentanol>cyclohexanol (Dpa(V) values were not given), and xylene, benzene, toluene and 1,2dichloroethane deviate from the line [88]. 3. Nature of the Extracted Complexes and Extraction Enthalpy The slope of the log DU(VI) vs. log aHCl dependence has been found to be 3.2 with DPSO [94] and 2.0 or 2.2 with DOSO [2, 94], both in CCl4. Thus, U(VI) has been suggested to be extracted as a solvated trichloride species by DPSO but as solvated dichloride by DOSO [94]. Trisolvates are the most frequently reported extracted

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

199

species but, in some cases, different extracted species have been found by different authors even in identical systems. To summarize the reports, the species extracted are • • • • • • •

UO2Cl2 · 2HCl · 2B by DPSO in CCl4 from 3 M HCl [94] UO2Cl2 · 3B together with UO2Cl2 · HCl · 3B by DPSO in CCl4 from 3 M HCl [88] UO2Cl2 · 3B by DHxSO and DHpSO in CCl4 from 3 M HCl [88] UO2Cl2 · 3B by DOSO in CCl4 from 3 M HCl [88, 94] UO2Cl2 ·3B by DOSO in CCl4, heptane, and benzene from weakly acidic solutions of UO2Cl2 [87] UO2Cl2 · 2B by DOSO in chloroform from weakly acidic solutions of UO2Cl2 [87] UO2Cl 2 · 3B by DNSO and DDSO in CCl4 from 3 and 5 M HCl [88], respectively

Polymerized complexes at >0.006 M U(VI) in the organic phase are presumed to be extracted by 0.5 M DPSO in CCl4 from 3 M HCl [94]. The Cr(VI) complex formed in the extraction by DBSO in benzene possibly embodies a chloride ion [50]. The log Dpa(V) vs. log aHCl dependence in 5 M HCl+variable LiCl has a slope of ~3 with DPSO in CCl4 and DΦSO in benzene [92]. Probably solvated trichloride species such as PaOCl3 or Pa(OH)2Cl3 are extracted, even if a Dpa(v) value based on the thermodynamic activity of Pa(V) in the aqueous phase should be plotted vs. aHCl. A trisolvate PaOCl3 · 3B is extracted by DOSO [88, 92] and DΦSO [92] in CCl4 from 7 or 8 M HCl. DPSO in CCl4 is reported to extract a mixture of a disolvate and a trisolvate from 8 M HCl [92] or only a trisolvate from 7 M HCl [88]. It has been substantiated by infrared spectra that the uranyl ion is bound to the O atom of DOSO. The S→O frequency is 1051 cm-1 in the spectrum of solid UO2C12 · 2B, but it is 925 cm-1 in pure DOSO [2]. Enthalpy of the extraction of a small amount of U(VI) by 0.7 M DOSO in xylene from 3.0 M HCl is -13.4 kJ mol-1 [94].

C. Tetravalent Metals 1. Distribution Data A survey of important sources of distribution data is given in Table 9. As implied by the table, rather little interest has been paid to the subject. Th(IV) is weakly extractable. Distribution ratios of 0.007–0.037 have been observed in the extraction of Th(IV) by 30 vol% DBSO in xylene from 1–9 M HCl [74]. In the extraction by 0.1 M DOSO in chloroform, the DTh value is 8 M HCl [13]. The rate of the extraction of Pt(IV) by 0.4 M PetrSO in kerosene from 2 to 6 M chloride ions was studied in a stirred cell at 4–50°C. The rate is controlled by the

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

201

diffusion of the complex H2PtCl6 · 2B · 6H2O from the interface into the bulk organic phase [101]. 2. Effect of Extraction Variables There is an appreciable effect of the extractant structure on the extraction of Pt(IV). Branching suppresses the extraction efficiency in the order DOSO>DEHSO>di(2octyl) sulfoxide>di(3-octyl) sulfoxide, and the introduction of a ring substituent suppresses it in the order DHxSO>cyclohexyl hexyl sulfoxide>dicyclohexyl sulfoxideⰇDΦSO, both orders observed at 6 M HCl [100]. The structure effect is further illustrated by extensive results by DHxSO and asymmetrical sulfoxides (see Fig. 11). The effect of branching is not clear with butyl branched-alkyl sulfoxides, where the extraction ability decreases in the order of the varied alkyl group 2-ethylhexyl>3,5,5-trimethylhexyl>octyl. Notice that at the same number of C atoms DHxSO is more effective than butyl octyl sulfoxide. With 2-tolyl as the invariant group, the extraction ability decreases with increasing branching in the varied group order octyl>2-ethylhexyl>3,5,5-trimethylhexyl. Replacement of butyl as the invariant group by 2-tolyl suppresses the extraction ability much more than any variation of an alkyl group. TBP extracts Th(IV) chloride less effectively than DPSO and DOSO, but slightly more effectively than DΦSO (CCl4 diluent) [102]. The efficiency of the Pt(IV) extraction from 1–6 M HCl decreases in the sequence tributylphosphine oxide>Nbutyl octanamide>DHxSO>dibutyl butylphosphonate>N,N-dibutyl octanamide [100].

Figure 11 Extraction of Pt(IV) by 1 M DHxSO and asymmetrical sulfoxides in xylene. Initially, 0.005 M Pt(IV) in the aqueous phase, room temperature. (From Ref. 99.)

Copyright © 2004 by Marcel Dekker, Inc.

202

Kolarik

The efficiency of the Th(IV) extraction by 0.25 M DPSO from 7 M HCl decreases in the diluent (DTh) order CS2 (0.34)>CCl4=cyclohexane=benzene (all 0.29)>toluene (0.22)~chlorobenzene (0.21)>1,2-dichloroethane (0.086)>1-pentanol (0.033). Log DTh was suggested to be a linear logarithmic function of the diluent dielectric constant [102], but the plot is not very convincing. In the extraction of Zr(IV) by 0.25 M DPSO from 7 M HCl the sequence is benzene (34.0)>TCE (31.3)>chloroform (7.2)>benzonitrile (1.92)>1-pentanol (1.53) [97]. Finally, in the extraction of Hf(IV) by 0.1 M DPSO from 7 M HCl the diluent order is nitrobenzene (DHf=121)>benzene>toluene>xylene>CCl4>chloroform>m-cresol (DHf=0.002) [103]. 3. Nature of Extracted Complexes and Extraction Enthalpy Th(IV) was said to be extracted by DPSO as a pentachloro species, but as a tetrachloro species by DOSO. This was based on a rather ambiguous evaluation of curved log DTh vs. log aHCl dependencies. The limiting slopes have been interpreted as 5.0 with DPSO and as 4.0 with DOSO, both in CCl4, although they could be 4 and 3, respectively, as well. The slopes of the logarithmic dependencies of DTh on the activities of the Cl- and H+ ions (assuming them to be separable) in HCl+NH4Cl mixtures are 5.0 and 1.16, respectively, with DPSO. Thus, a pentachloro and a tetrachloro species have been suggested to be extracted by DPSO and DOSO, respectively [102]. On the other hand, it was concluded that Hf(IV) was extracted as a tetrachloro species by DPSO in CCl4, because the log DHf vs. log aHCl dependence had a slope of 4.0 [103]. Mono- to trisolvates were reported to be extracted, namely, • • • • • • • • •

ThCl4 · HC1 · 2B by DPSO in CCl4 from 7 M HCl [102] ThCl4 · 2B by DOSO in CCl4 from 7 M HCl [102] ThCl4 · 3B by DΦSO in benzene from 7 M HCl [102] UCl4 · HC1 · 3B by DOSO in toluene from 7 M HCl and 1 M HCl+6 M LiCl [13] ZrCl4 · 2B by DBSO in benzene from 7 M HCl [50] ZrCl4 · 2B and ZrCl3(ClO4) · 2B by DPSO in CCl4 from unclearly specified aqueous solutions [104] ZrCl4 · B by DPSO and DOSO in TCE from 7 M HCl [97] HfCl4 · 3B by DPSO and DOSO in CCl4 from 6 M HCl [82] TeCl4 · HCl · 2B by DHxSO in xylene from 4 M HCl [75]

Pt(IV) and Ir(IV) are suggested to be extracted as the ion pairs [B2 · H3O+]2 and The ion-pair nature of the species was implied by the fast extraction rate and the presence of the hexachloro anions is indicated by near uv and visible absorption spectra, but the presence of a water molecule was not proved.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

203

In the extraction of U(IV) by 0.05 M DOSO in toluene from 7 M HCl the DU(IV) value decreases with increasing temperature at 5–50°C [13]. In the extraction of Th(IV) by DPSO in CCl4 from probably 7 M HCl the log DTh vs.1/T dependence is curved at 10–50°C and a rough estimate of ∆H ~13 kJ mol-1 has been made [102]. In the extraction of Hf(IV) by 0.1 M DPSO or DOSO in CCl4 from 7 M HCl, DHf increases with the temperature and ∆H is 54.2 and 55.3 kJ mol-1, respectively [103].

D. Other Metals 1. Distribution Data and Extraction Kinetics Selected distribution data are surveyed in Table 10. Notice that also extraction of those elements has been studied which prefer to be bound to soft donor atoms. Weakly or negligibly extractable are • • • •

• • •

Eu(III) by 10% PetrSO in toluene from 0.1 to 8 M HCl (DEu(III)≤0.004) [90] and by 0.1 M methyl 4,8-dimethylnonyl sulfoxide in xylene from 7 to 11 M HCl (DEu(III)DHxSO at 4–6 M HCl [99], and DHxSO~

Figure 13 Extraction of selected transition metals by 10% PetrSO in toluene. Trace metals, room temperature. (From Ref. 90.)

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

207

Figure 14 Extraction of Hg(II) by 0.5 M DEHSO in xylene (upper curve) [107] and by 0.2 M DHxSO in benzene [20], and extraction of Zn(II) and Cd(II) by 0.2 M DHxSO in benzene [106]. Initially, ≤0.0037 M Hg(II), 0.0073 M Zn(II), and 0.005 M Cd(II), 25°C.

DOSO>DDSO at 2–6 M HCl [100]. At a constant total number of C atoms, DHxSO ~butyl octyl sulfoxide at 0.2–2 M HCl but DHxSO>butyl octyl sulfoxide at 3–6 M HCl [99]. Branching suppresses the extraction ability for Pd(II) in the order DOSO>DEHSO>di(2-octyl) sulfoxide>di(3-octyl) sulfoxide, and the ring effect is illustrated by the order DHxSO>cyclohexyl hexyl sulfoxide>dicyclohexyl sulfoxideⰇDΦSO, both orders at 6 M HCl [100]. The structure effect in the extraction of Au(III) and Pd(II) by asymmetrical sulfoxides is illustrated in Table 11. Notice there that the invariant substituent influences the effect of the alkyl branching and can change the effect of the replacing of butyl by 2-tolyl. 0.1 M xylene solutions of methyl 4,8-dimethylnonyl and methyl 3-ethylheptyl sulfoxides (both at 1–10 M HCl) and of methyl pentadecyl sulfoxide (at 1–4 M HCl) exhibit similar extraction ability for Fe(III) [8]. The extraction efficiency of Hg(II) from 0.1 M HCl decreases in the order DOSO>DEHSOⰇDΦSO, all in xylene [107]. Surprising is the behavior of Pd(II). It is highly extractable by PetrSO from 0.1 to 8 M HCl (Fig. 13), while in the extraction by 0.2 M DEHSO in toluene the Dpd(II) value decreases from 16.4 at 0.01 M HCl to 0.4 at 1–5 M HCl [85]. To compare sulfoxides with other solvating extractants, the ability to extract CoCl2 as a disolvate into heptane, benzene or chloroform increases in the series TBP< DOSOchloroform (0.04) [107]. 3. Nature of Extracted Complexes and Extraction Enthalpy The extracted species can be mono- to trisolvates of neutral complexes or complex acids, namely, • • • • • • •

InCl3 · 2B · H2O at 2 M HC1 [105] Trisolvates of In(III) and Ga(III) chlorides by DHxSO in toluene from 4 M HCl [90] Possibly a mixture of di- and trisolvates of Fe(III) chloride by DHxSO in toluene from 4 M HCl [90] HFeCl4 · 2B and HFeCl4 · 2B at >4 M HCl and FeCl3 · nB (with probably n=2 and 3) by methyl 4,8-dimethylnonyl sulfoxide in xylene from ≤1 M HCl [8] MCl2 · 2B, HMCl3 · 3B, and H2MCl4 · 2B with M=Zn, or Cd by DHxSO in benzene from 7 M HCl [106] HgCl2 · B, HgCl2 · 2B, HHgCl3 · 3B, and H2HgCl4 · 2B by DHxSO in benzene from 0.01–1 M HCl [20] HgCl2 · 3B by DOSO, DEHSO, and DΦSO in xylene from 0.1 M HCl [107]

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists • •

209

Chloride bridged [PdCl2 · B]2 by PetrSO in kerosene from 0.1 to 6 M HCl [108] PdCl2 · 2B by DHxSO in xylene from 2 M HCl [100]

It was suggested that DBSO in benzene extracted Fe(III) as the species [50], but this was a speculative claim. The water concentration in the organic phase was not measured, nor was the degree of the chloride complexation of Fe(III) in the aqueous phase assessed. Also speculative is the suggestion that the ion pairs and are extracted by DHxSO in xylene [100]. The presence of the tetra- and pentachloro anions in the species was assumed because characteristics of the hexachloro anions were missing in absorption spectra, and the presence of the protonated water molecule was not at all proven. The S→O stretching vibration of DHxSO is shifted from 1050 to 990 cm-1 when DHxSO is complexed by Hg(II) in benzene [20]. In the extraction by DHxSO in benzene ∆H=-20.3 and -35.2 kJ mol-1 was found for Hg(II) at 0.01 and 6 M HCl respectively [20], and for Zn(II) ∆H=-36.9 kJ mol-1 [106] and for Cd(II) ∆H=-38.4 kJ mol-1 were found at 6 M HCl [106].

IV. EXTRACTION FROM OTHER MEDIA A. Perchloric, Pertechnetic, and Perrhenic Acids Perchloric acid is less extractable than nitric acid, but substantially more extractable than hydrochloric acid. The concentration of HClO4 in 0.02 M ethyl dodecyl sulfoxide in xylene is 0.01 M after contact with 6 M aqueous acid, while the HNO3 and HCl concentrations are 0.02 and ~3×10-4 M, respectively [3]. Although DBSO in benzene appears to extract a monosolvate of HClO4 [14], the suggestion that the extracted species is the ion-pair with a DBSO molecule bound to the H+ ion via a water molecule is speculative. The amount of HClO4 extracted by DBSO cannot be assessed, due to unclear definition of the scales of figures in [14]. Perrhenic acid is well extractable by 0.7 M sulfoxides in xylene from 1 M H2SO4. With DHxSO the organic concentration of HReO4 is 0.07–0.25 M at 0.03–0.40 M Re(VII) in the aqueous phase. With 2-heptylthiophane S-oxide it is 0.11–0.32 M at 0.017–0.68 M Re(VII) in the aqueous phase [27]. Pertechnetic acid is not extracted by DBSO in benzene from 5×10-4 M NH4TcO4 containing an equivalent amount or excess (1 M) HCl [14].

B. Bivalent to Hexavalent Metals 1. Distribution Data A survey of selected distribution data is given in Table 12. U(VI) has been the most studied extracted metal and thiocyanate media have been the preferred ones.

Copyright © 2004 by Marcel Dekker, Inc.

210

Kolarik

Table 12 Survey of Data on the Distribution of Miscellaneous Metals Between Solutions of Common Sulfoxides and Aqueous Solutionsa

a

Single temperature value applies to the measurement of isothermal concentration dependencies of DM, a temperature range to the measurement of temperature dependency. rt: room temperature.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

211

2. Effect of Extraction Variables Perchloric and phosphoric acids can be used as salting-out agents in the extraction of perchlorates and phosphates. U(VI) perchlorate is extracted rather effectively. In the extraction by 0.5 M DOSO in 1,2-dichloroethane the DU(VI) vs. [HClO4] dependence passes through a maximum, rising from DU(VI)=0.9 at 0.05 M HClO4 to DU(VI)=110 at ~1.5 M HClO4 and falling to DU(VI)=9 at 4.5 M HClO4. Still higher efficiency is attained from 0.025 M HClO4+0.05 to 4.2 M LiClO4, with DU(VI)=1400 at 2.0 M LiClO. HClO4 competes even at low concentration. In the extraction from 1.0 M (Li,H)ClO4 the DU(VI) value sinks from 1900 at pH 2–3 to 35 at pH 0 [116]. The extraction of U(VI) phosphate by 0.1 M DPSO [115] and 0.32 M PetrSO [46, 47] in benzene from 1 to 6 M H3PO4 yields DU(VI) values increasing from 0.6 to 3.9 and from 0.4 to 1.8, respectively. Thiocyanates and picrates are not extractable from strongly acidic solutions and, as seen in Table 12, their distribution has been studied at pH 3. The extraction of Tm(III) thiocyanate becomes less effective in the order DOSO>DPSOⰇDΦSO in benzene [119], but that of Yb(III) thiocyanate varies in the order DPSO>DOSO [120]. In the extraction of Hg(II) iodide by DEHSO in benzene the DHg(II) value decreases from 2300 at 1×10-4 M I- to 2.1 at 0.04 M I-. The effect is ascribed to the formation of anionic triiodo and tetraiodo complexes in the aqueous phase [125]. Similar decrease of the DHg(II) value was observed and analogously explained in the extraction of Hg(II) chloride (see Fig. 14). 1,2-Bis(octylsulfinyl)ethane in butyl acetate extracts Pt(IV) and Pd(II) weakly from 2 to 6 M HCl solutions. Addition of potassium iodide strongly enhances the extraction of Pt(IV) (under simultaneous reduction to Pt(II)) [121] and Pd(II) [122]. Dialkyl sulfoxides are weaker extractants than trioctylphosphine oxide with its strongly basic phosphoryl oxygen atom. This applies to the extraction of Tb(III) [113] and Yb(III) [120] thiocyanates by DPSO and DOSO in CCl4, and of Cd(II) and Hg(II) iodides by DEHSO in benzene [125]. The diluent effect was investigated in the extraction by 0.3 M DPSO from 1 M NH4SCN at pH 3. For Yb(III) the diluent order is benzene (DYb=0.72)>xylene> CCl4>benzonitrile>chlorobenzene>m-cresol>chloroform (DYb=0.001) [120]. Tb(III) behaves similarly, and the diluent (DTb(III)) order is benzene (0.43)>xylene (0.36)>CCl4 (0.24)>chlorobenzene (0.17)>benzonitrile (0.15)>m-cresol (0.04) [113]. A somewhat different diluent (DCe(III)) order was reported for Ce(III), namely, CCl4 (0.50)>xylene (0.346)>benzene (0.295)>benzonitrile (0.181)>chlorobenzene (0.102)>m-cresol (0.020)>chloroform (0.0007) [112]. Again a different order of the diluent-(DU(VI)) pairs was found in the extraction of 0.01 M U(VI) by 0.2 M DOSO from 1 M HClO4. It is nitrobenzene (19.0)>TBP (8.1)>1.2-dichloroethane=2-octanone (3.0)>butyronitrile (2.33)>MiBK (1.86)> 1butanol (1.22)>1-octanol=chloroform=chlorobenzene (1.00). A flaky or oily third phase is formed with butyl acetate, diethyl ether, benzene, toluene, xylene, CCl4, cyclohexane, and methyllaurate as diluents [116]. The efficiency of the Am(III)

Copyright © 2004 by Marcel Dekker, Inc.

212

Kolarik

extraction by 0.1 M DOSO from 0.005 M picrate at pH 3.0 decreases in the diluent (DAm(III)) order nitrobenzene (284)>toluene (173)>CCl4 (76.7)>dodecane (14.7) ~dichloromethane (14.4)>chloroform (0.23) [117]. 3. Nature of Extracted Complexes and Thermodynamic Functions The solvation number of thiocyanates shows a weak regularity with regard to the charge of the extracted ion. The complexes formed in the organic phase are • • • • • • • • • • •

UO2(SCN)2 · 2B with DPSO and UO2(SCN)2 · 3B with DOSO, both in chloroform, from 1 M NH4SCN at pH~3 [111] Th(SCN)4 · 3B with DPSO and DOSO in chloroform from 1 M NH4SCN at pH~3 [111] Am(SCN)3 · 4B with DPSO, DOSO in benzene from 1 M NaSCN, and Am(SCN)3 · 5B with DΦSO from 2 M NaSCN [123] Ce(SCN)3 · 4B with DPSO and DOSO in CCl4 from 1 M NH4SCN at pH 3 [112] Nd(SCN)3 · 4B with DEHSO and DOSO in benzene from 1 M NH4SCN at pH 3 [124] Eu(SCN)3 · 4B with DEHSO and DOSO in benzene from 1 M NH4SCN at pH 3 [124] Tb(SCN)3 · 4B with DPSO and DOSO in CCl4 from 1 M NH4SCN at pH 3 [113] Predominantly Er(SCN)3 · 3B with DEHSO and DOSO in benzene from 1 M NH4SCN at pH 3 [124] Tm(SCN)3 · 4B with DPSO, DOSO, and DΦSO in benzene from 3 M NaSCN at pH 3 [119] Yb(SCN)3 · 4B with DPSO and DOSO in CCl4 from 1 M NH4SCN at pH 3 [120] Zn(SCN)2 · 2B and Cd(SCN)2·4B with DEHSO in benzene from 0.1 M NH4SCN at pH 3 [114]

No regularity is revealed in other media, due to the diversity of the data. The extracted complexes are • • • • •

UO2(ClO4)2 · 4B with DOSO in 1,2-dichloroethane from 1 M HClO4 [116] Hf(ClO4)4 · 2B with DOSO in CCl4 from 6 M HClO4 [82] HNbOF · 3B with PetrSO, DHxSO, and 2-heptylthiophane in xylene from 5 M HF+5 M H2SO4 [118] HTaF6 · 3B or H2TaF7 · 3B with PetrSO, DHxSO, and 2-heptylthiophane Soxide in xylene from 5 M HF+2.5 M H2SO4 [118] ZnI2 · 2B with DEHSO in benzene from 0.01 M I- at pH 3 [125]

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

213

Table 13 Thermodynamic Values of the Extraction of Trace Metals in Miscellaneous Systems

• • •

HgI2, HgI2 · B, and HgI2 · 2B with DEHSO in benzene from 0.01 M I- at pH 3 [125] UO2A2 · 2B and AmA3 · 2B with A=picrate and B=DOSO in chloroform from 0.01 M picrate at pH 3.0 [117] PtI2 · B [121] and PdI2 · B [122] with 1,2-bis(octylsulfinyl)ethane in butyl acetate from 2 to 6 M HCl+0.1% KI.

It has been concluded from infrared spectra that complexed Pt2+ [121] and Pd2+ [122] ions are bound to sulfur atoms of 1,2-bis(octylsulfinyl)ethane. Thermodynamic functions of the extraction equilibria (see Table 13) were determined in picrate and thiocyanate systems, in each of them with the same diluent and at similar concentration ranges. This makes them satisfactorily comparable.

V. SELECTIVITY OF THE EXTRACTION Of general interest is discrimination within pairs or groups of chemically similar elements, such as lanthanides(III), actinides(III), Zr(IV)–Hf(IV), or Nb(V)-Ta(V). Of special interest is selectivity for particular elements to be separated and purified in a solvent extraction process. For example, selectivity for U(VI) and Pu(IV) over fission products is needed in the reprocessing of nuclear fuel and selectivity for transplutonides(III) over lanthanides(III) would be needed in the partitioning of nuclear wastes. Sequences of the extractability of a series of elements are illustrated in Figs. 5 and 7 for nitrate systems, and in Fig. 10 (curves 3 and 4) and Figs. 12–14 for

Copyright © 2004 by Marcel Dekker, Inc.

214

Kolarik

chloride systems. The figures show nitric and hydrochloric acid dependences of DM and can thus accommodate data on a limited number of elements only. Data on as many as 14 elements are shown in Fig. 15, indeed at one or two acid concentrations for each extractant. Not shown are data on 14 elements from Ref. 79 which are inconsistent within the paper itself and with data in Ref. 74. Still more extensive data on a perchlorate system are given in Table 14. It is quite obvious that no general characterization of the selectivity of sulfoxides can be given. Due to the different forms of the D M vs. [acid] dependencies, the discrimination between extracted elements is strongly dependent on the acid concentration (Fig. 16) and extractant structure. Moreover, discrimination between elements must be dependent on the concentration of the extractant if they are extracted in the form of different solvates. Finally, the temperature can also play a role. Striking differences are reported between the U(VI)/Th(IV) separation factors αU(VI)/Th obtained with different sulfoxides at two HNO3 concentrations (1 M extractant in xylene, 0.2 M U(VI) or Th(IV)). At 1 M HNO3 the extractant (αU(VI)/ Th/DU(VI)) order is 2-ethylhexyl p-tolyl sulfoxide (15.5/1.46)>di(2-octyl) sulfoxide

Figure 15 Sequences of the extractability of elements in various nitrate systems. Curve 1— 30 vol% DBSO in xylene/2 M HNO3 [74]; 2—30 vol% DBSO in xylene/7 M HNO3 [74]; 3—50% DBSO in cyclohexane/8 M HNO3 [81]; 4—1 M DHpSO in TCE/6M HNO3 [34]. Initially, ≤0.01 M metals, room temperature (systems 1–3) or 20°C (system 4).

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

215

Table 14 Extraction of Elements by 0.5 M DOSO in 1,2-Dichloroethane from 0.025 M HClO4+1.0 M LiClO4a

Initially 0.01 M elements, room temperature. Slow reduction to Ce(III). c HCl required to keep Sn(IV) in solution. Source: Ref. 116. a

b

Figure 16 Separation factors in the extraction by 1 M DHpSO in TCE. Initially, ⱕ0.01 M metals, 20 °C. (From Ref. 34.)

Copyright © 2004 by Marcel Dekker, Inc.

216

Kolarik

Table 15 Separation of U(VI) and Th(IV) in the Extraction by DPSO and DOSO in CCl4a

0.25 M extractants, 0.02 M U(VI) or Th(IV), room temperature. Source: Ref. 102.

a

(9.3/1.12)>dodecyl p-tolyl sulfoxide (4.0/1.08) ~DEHSO (3.96/2.65)>di (pethylphenyl) sulfoxide (0.54/0.24). At 4 M HNO3 the sequence is di(p-ethylphenyl) sulfoxide (12.1/1.21)>2-ethylhexyl p-tolyl sulfoxide (7.47/4.41)>dodecyl p-tolyl sulfoxide (2.4/2.36) ~DEHSO (2.4/3.16)>di(2-octyl) sulfoxide (0.010/0.20) [11]. Again for separation of the U(VI)–Th(IV) pair, another example of the effect of the extractant structure and acid concentrations is given in Table 15. Temperature can also influence the discrimination between two extracted metals. Table 16 shows it for the extraction of Pu(IV) and U(VI) by DHxSO and DOSO. As seen in Fig. 8, the separation potential of sulfoxides within the lanthanide(III) group is limited. Practicable separation factors between two adjacent lanthanides(III) are attained only for the La(III)-Ce(III) and, to a lesser extent, Ce(III)-Pr(III) pairs. This was also observed in the extraction by 0.1 M DOSO in benzene from an unspecified aqueous solution [84].

Table 16 Comparative Extractability of Pu(IV) and U(VI) in the Extraction by DHxSO and DOSO in Solvesso 100a

0.2 M extractants, trace Pu(IV) and U(VI), 2 M HNO3. Source: Ref. 72. a

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

217

No unique value for the Zr(IV)/Hf(IV) separation factor can be determined from the data on the extraction by 1 M DHpSO in TCE from 7 to 10 M HNO3 [34]. The reason is that the DHf(IV) value decreases with increasing Hf(IV) concentration even at very low solvent loading. Typically the ␣Zr/Hf value is >10. The extractability of platinum metals by 0.5 M DHxSO in xylene from 1 to 6 M HCl decreases in the order Pd(II)>Pt(II)>>Ir(IV)>Pt(IV)>Rh(III)>Ir(III) [100]. The experiments were performed at 40°C with Pt(II), Rh(III), and Ir(III) and at 20°C with the other platinoids, but this need not diminish the plausibility of the order. The selectivity of the synergistic extraction by combinations of acidic chelating and sulfoxide extractants is dealt with in Chapter VIIB.

VI. INTERFERING PHENOMENA A. Third-Phase Formation The solubility of some extracted complexes in the organic phase is so limited that, if a high solvent loading is attained, they separate in the form of a third phase. In some cases the third phase is a solid, but in many systems it is a liquid which forms a second, heavy organic phase. It contains a considerable fraction of the extracted metal, and it causes serious disturbances in countercurrent operations and also batch distributions on any scale. For example, a third phase was reported to be formed in the extraction of 0.01 M U(VI) from 1 M HClO4 by 0.1 DOSO in butyl acetate, diethyl ether, benzene, toluene, xylene, CCl 4 , cyclohexane and methyl laurate, but without a characterization of its properties [116]. Obviously the perchlorate system is highly sensitive to the third phase formation. In nitrate systems the splitting of the organic phase into two layers is limited to solutions of sulfoxides in nonpolar, mainly aliphatic diluents. A short mention in Ref. 77 can be understood in the manner that U(VI) forms a third phase if its concentration in 1.1 M DEHSO in dodecane exceeds 0.21 M, when extracted from aqueous 2 M HNO3. The formation of the third phase was studied extensively in the extraction of U(VI) and Pu(IV) by DEHSO in dodecane from HNO3 solutions [62]. The accuracy of otherwise interesting data may be somewhat impaired by insufficient constancy of the temperature (±2°C), because the formation of a third phase is generally a temperature-sensitive phenomenon. The limiting organic concentration (LOC) of the extracted metal, i.e., the maximum attainable concentration at which still no third phase is formed, is a function of the extracted metal and the composition of the phases. Figures 17 and 18 show that the U(VI) complex is much more soluble than the Pu(IV) complex, and the LOC of both metals slightly decreases with increasing HNO3 concentrations. The LOC of Pu(IV) is suppressed by a factor of ~1.3 if 0.4 M DEHSO is at 1–2.5 M HNO3 loaded with U(VI) to a concentration of 0.0084 M.

Copyright © 2004 by Marcel Dekker, Inc.

218

Kolarik

Figure 17 Limiting organic concentration of U(VI) as a function of the aqueous HNO3 concentration at (25±2)°C. (From Ref. 62.)

The LOC increase with the DEHSO concentration (Figs. 17 and 18) can be ascribed to an enhancement of the polarity of the organic phase. Also, alcohols increase the LOC. Figure 19 shows this for 2-ethyl-1-hexanol, and isodecanol gives a similar picture. It should be remembered that alcohols suppress the extraction. Inert salts not only support the extraction, they also enhance the LOC. Addition of 2 MLiNO3 or 0.5 M Ca(NO3)2 to 2 M HNO3 increases the LOC of U(VI) and Pu(IV) by a factor of ~1.3 and ~1.2, respectively [62].

Figure 18 Limiting organic concentration of Pu(IV) as a function of the aqueous HNO3 concentration at (25±2)°C (From Ref. 62.)

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

219

Figure 19 Limiting organic concentrations of U(IV) and Pu(IV) as a function of the HNO3 concentration at various concentrations of 2-ethyl-1-hexanol (EHOH). Open points: U(VI), solid points: Pu(IV), 0.4 M DEHSO in dodecane, (25±2)°C. (From Ref. 62.)

Data on Pu(IV), as given in Ref. 62, have been described by the empirical equation

where S is the LOC and S° is the hypothetical solubility of the extracted Pu(IV) complex in 0.4 M BEHSO in dodecane at and 298 K. ϕm is the volume fraction of diluent modifier, S and S° are in mmol/L, and the other concentrations are molarities. The Sechenov parameters were estimated as S°= 10.4 mmol/L, ka=0.14, kU=2.97×10-5, km=121 for 2-ethyl-1-hexanol and km=71.9 for isodecanol, and ks=-0.166 for LiNO3 and ks=-0.427 for Ca(NO3)2 [127].

B. Radiation Damage Knowledge of the radiation stability of sulfoxides is of importance, due to their intended use in nuclear processing. The radiation stability has typically been studied with phases irradiated by gamma rays of a 60Co source. Unfortunately, the conditions of the irradiation are seldom sufficiently described. Often missing is information about temperature and the presence of air, and it can only be assumed that the phases were not stirred during the irradiation. Figure 20 shows the change of distribution ratios caused by radiation damage of DEHSO. The DU(VI) and Dpu(IV) values decrease up to a dose of ~18 MRad, mainly due to the lowering of the concentration of undamaged DEHSO. A sulfone is indicated by infrared spectra to be the main degradation product, and it seems not

Copyright © 2004 by Marcel Dekker, Inc.

220

Kolarik

Figure 20 Effect of irradiation on the extraction of U(VI), Pu(IV), Zr(IV), and fission products (F.P.) by DEHSO. Open points: 0.4 M DEHSO (U, Pu) or 0.2 M DEHSO (Zr, F.P.) in dodecane from 2 M HNO3 [129]; solid points: 0.2 M DEHSO in kerosene from 4 M HNO3 [11]. Real mixture of fission products used, room temperature.

to extract U(VI) and Pu(VI) noticeably. The increase of DU(VI) and DPu(IV) at >18 MRad is ascribed to unidentified products of radiolysis and hydrolysis, which possibly act as synergists in the presence of DEHSO [129]. An unlike behavior was observed in systems involving Solvesso 100 diluent and 2 M HNO3. In the extraction by 0.2 M DHxSO or DOSO, DU(VI) decreases continuously from 2.3 to 0.7 when the radiation dose increases from zero to 169 MRad [128]. With 0.2 M DHxSO, Dpu(IV) increases from 1.10 before irradiation to 2.44 at 1.24 MRad and to 3.25 at 8.7 MRad [26]. With 0.2 M DOSO, DPu(IV) changes from 5.1 before irradiation to 6.8 at 8.5 MRad and to 5.7 at 36.7 MRad [38]. The extraction of Zr(IV) from 2 M HNO3 is enhanced by the radiolysis of DEHSO in kerosene [11] and dodecane [129] (Fig. 20) as well as of DHxSO [26, 128] and DOSO [38, 128] in Solvesso 100. DEu(III) is 2.2×10-4, 3.0×10-4, and 5.2× 10-4 at 0, 1.24, and 8.7 MRad, respectively, in the extraction by 0.2 M DHxSO in Solvesso 100 [26]. Data on Ru(III) [11, 26, 128] were given for an unspecified trivalent form, but more useful would be data on nitrosylruthenium(III) which is the typical form in the reprocessing of nuclear fuel. Figure 20 shows the extraction of the sum of real fission products as a function of the DEHSO degradation. A comparison with TBP is appropriate. Only one source [11] compares directly the effect of radiation damage on the extractant properties of a sulfoxide and TBP. Data excerpted from the source are given in Table 17. They clearly show that the U(VI)/ Zr(IV) separation is deteriorated by radiation damage much more with TBP than with

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

221

Table 17 Comparative Effect of Radiation on the Extraction of U(VI) and Zr(IV) by DEHSO and TBP. 0.2 M Extractants in Kerosenea

a Initially 0.05 M U(VI) and trace Zr(IV) in the aqueous phase, probably room temperature. Source: Ref. 11.

DEHSO. It is favorable that the comparison was made with the kerosene diluent which is preferred in nuclear processes.

VII. SYNERGISM A. Solvating/Solvating and Solvating/Basic Combinations of Extractants These extractant combinations have been studied rather extensively although, as experience shows, only moderate synergistic enhancement can be expected. Data in Table 18 corroborate the expectation. Synergistic enhancement (SE) of >10 has been observed only in one case and even strong antagonismus is not unusual. Extensive data on the extraction of U(VI) by the B1–B2 mixtures DPSO-DOSO, DPSO-D ΦSO, and DPSO-TBP show that the SE value depends on the concentrations of the two extractants, attaining a maximum or changing monotonously within the studied limits. Table 18 shows the highest and lowest SE values, found at a constant sum of component concentrations. A variety of synergistic complexes were reported to be extracted, namely, • • • • • • •

UO2(NO3)2 · B1 · B2 by DOSO and di(2-octyl) methylphosphonate in benzene from 1 M HNO3 [63] UO2(NO3)2 · B1 · B2 by DOSO and TBP in CCl4 from 1 M HNO3+1 M NaNO3 [36] UO2(NO3)2 · B1 · B2 by DΦSO and TBP in benzene from 4.2 M HNO3 [66] UO2(NO3)2 · B1 · B2 by PetrSO and octyl decyl sulfoxide in CCl4 from 2 M HNO3 [48] UO2(NO3)2 · B1 · B2 by PetrSO and TBP in kerosene from 2 M HNO3 [44] UO2(NO3)2 · B1 · B2 by PetrSO and TBP in benzene from 2 M HNO3 [67] Ln(SCN)3 · xB1 · (4–x)B2 with Ln=Nd and Eu, by B1=DOSO and B2=DEHSO or TOPO in benzene from 1 M SCN- at pH 3 [124]

Copyright © 2004 by Marcel Dekker, Inc.

222

Kolarik

Table 18 Synergistic Enhancement in the Extraction by Solvating/Solvating and a Solvating/ Basic Mixturesa

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists Table 18 Continued

Only the highest or both the highest and the lowest reported SE values are given for each system. Di(2-octyl) methylphosphonate. c Octyl decyl sulfoxide. d D1 and D2 are not given in the original source. e Aliquat 336. f Alamine 336. a

b

Copyright © 2004 by Marcel Dekker, Inc.

223

224

Kolarik

Table 19 Survey of Data on the Synergistic Extraction by an Acidic Extractant and a Sulfoxide Synergista

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

225

Table 19 Continued

BMPPT: 4-benzoyl-2,4-dihydro-5-methyl-2-phenyl-3H-pyrazol-3-thione, Cyanex 272: bis(2,4, 4-trimethylpentyl)phosphinic acid, Cyanex 301: bis(2,4,4-trimethylpentyl)dithiophosphinic acid, DtBSA: 3, 5-di(tert-butyl)salicylic acid, HDEHP: di(2-ethylhexyl)phosphoric acid, HEh[EhP]: 2-ethylhexyl hydrogen 2-ethylhexylphosphonate, PBI: 3-phenyl-4-benzoyl-5-isoxazolone, PMCBP: 1-phenyl-3-methyl-4(2-chlorobenzoyl)-5-pyrazolone. a Single temperature value applies to the measurement of isothermal concentration dependencies of DM(IV), a temperature range to the measurement of temperature dependency rt, room temperature.

• • •

Er(SCN)3 · xB1 · (3–x)B2 by B1=DOSO and B2=DEHSO in benzene from 1 M SCN- at pH 3 [124] Er(SCN)3 · xB1 (3–x)B2 by B1=DOSO and B2=TOPO in benzene from 1 M SCN- at pH 3 [124] Yb(SCN)3 · xB1 · (3–x)B2=DPSO and B2=DOSO or TOPO in CCl4 from 1 M SCN- at pH 3 [120]

For some systems no unambiguous composition of the extracted synergistic complex can be found by the slope analysis, because the slope of the log DM vs. log [B1] dependence changes when measured at various constant concentrations of B2. Such phenomena were observed, e.g., in the extraction of U(VI) by the pairs DPSO-DΦSO and DPSO-TBP in CCl4 from 3 M HCl [94], of Np(IV) by DiPSODOSO and DHxSO-DOSO in Solvesso 100 from 2 M HNO3 [25], of Th(IV) by

Copyright © 2004 by Marcel Dekker, Inc.

226

Kolarik

DPSO-DOSO in CCl4 from 7 M HCl [102], and of Ce(III) by DPSO-DOSO in CCl4 from 1 M NH4SCN at pH3 [112]. The enthalpy of the extraction of U(VI) by two extractant combinations was determined. It is ∆H=-26.2 kJ mol-1 in the extraction by PetrSO and octyl decyl sulfoxide in CCl4 from 2 M HNO3 [48], and ∆H=-23.1 kJ mol-1 in the extraction by PetrSO and TBP in kerosene from 2 M HNO3 [44].

B. Acidic/Solvating Extractant Combinations 1. Distribution Data Selected sources of distribution data in synergistic systems are gathered in Table 19. The table shows that the synergistic action of sulfoxides has mostly been studied in combination with TTA and 5-pyrazolone derivatives as acidic extractants. Unlike systems involving two solvating extractants, combinations of an acidic and a solvating extractant can in particular systems give synergistic enhancement as high as several orders of magnitude. Examples are shown in Figs. 21 and 22. Similar synergistic action has been found, e.g., in the extraction of Tm(III) by TTA+DPSO in benzene [119]. Less marked effects (SE≤55) were observed in the extraction of Hf(IV) by TTA+DOSO in CCl4 from 6 M HClO4 [82]. Moderate synergistic effects (SE≤14.0) were found, e.g., in the extraction of Pa(V) by TTA+DPSO in benzene from 6 M HCl and of Zr(IV) by the same solvent from 2 M HCl+2 M NaCl [91].

Figure 21 Synergistic action of DOSO in the extraction by TTA in benzene. The constant sum of the TTA and DOSO concentrations is 0.01 M (Th) and 0.02 M (Np and Pu). Aqueous phase: 1.0 M HClO4, room temperature. (From Ref. 136.)

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

227

Figure 22 Synergistic action of DOSO in the extraction by TTA in benzene. Th(TV)–0.01 M TTA; Np(IV)–0.02 M TTA; Pu(IV)–0.01 M TTA (lower curve); and 0.0125 M TTA (upper curve). Numerals on the points with arrow are the DTh and DNp(IV) values without DOSO. Aqueous phase: 0.1 M HClO4 (Th), and 1.0 M HClO4 (other metals). Room temperature. (From Ref. 136.)

Weak synergistic actions (SE=2-4) are exhibited, e.g., by DOSO in the extraction of Th(IV) by 2-ethylhexyl hydrogen 2-ethylhexylphosphonate, Cyanex 272 [bis (2,4,4-trimethylpentyl)phosphinic acid], and Cyanex 301 [bis (2,4,4trimethylpentyl)dithiophosphinic acid] [137]. DNSO and especially DΦSO enhance very weakly the extraction of U(VI) by naphthenic acid in benzene [142], and there is no synergistic effect in the extraction of Hf(IV) by TTA+DPSO and TTA+DOSO in CCl4 from 6 M HCl [82]. An antagonistic effect is caused by DΦSO in the extraction of Nd(III) [143] and Pd(II) [144] by 1-phenyl-3-methyl-4-dichloroacetyl-5-pyrazolone in chloroform. It was ascribed to the formation of a 1:1 molecular complex of the acidic extractant with DΦSO. A special case is the extraction of Zr(IV) by TTA from a chloride solution, where DPSO not only increases equilibrium distribution ratios but also accelerates the extraction rate [104]. Sulfoxide synergism can improve but also deteriorate the discrimination between two similar elements. The Eu(III)/Nd(III) separation factor is 6.2 with 0.3 M 4,4,4trifluoro-1-phenyl-1,3-butanedione alone in benzene, and is enhanced to 53 by addition of 0.005 M DEHSO [145]. On the other side, αEu(III)/Nd(III)=10.9 obtained with 01.01 M 1-phenyl-3-methyl-4-trifluoroacetyl-5-pyrazolone alone in chloroform is lowered to 7.8 by addition of 0.005 M DEHSO [146]. It appears that, generally, sulfoxide synergists cause only small enhancements or suppressions of separation factors within the lanthanide(III) series. The values αLu(III)/Eu(III)=2.7

Copyright © 2004 by Marcel Dekker, Inc.

228

Kolarik

and 5.1 are obtained with 0.01 M 1-phenyl-3-methyl-4-trifluoroacetyl-5-pyrazolone alone and its mixture with 0.005 M DEHSO, respectively, in chloroform [146]. With 0.01 M 3-phenyl-4-benzoyl-5-isoxazolone in xylene αTm(III)/Tb(III) is 4.52 and is changed to 5.13 by 1×10-4 M DOSO, to 3.96 by 1×10-4 M DEHSO, and to 2.63 by 0.001 M DΦSO. αTb(III)/Nd(III) is 15.5 and is changed to 18.1 by 1×10-4 M DOSO, to 19.3 by 1×10-4 M DEHSO, and to 16.2 by 0.001 M DΦSO [147]. Am(III) is separated with noticeable efficiency from unspecified lanthanides (III), if it is extracted by synergistic combinations of 0.1 M 4-benzoyl-2,4-dihydro5-methyl-2-phenyl-3H-pyrazol-3-thione in toluene from 0.1 M NaNO3 at pH 3.6. The αAm(III)/Ln(III) value is ≥28 in the presence of 0.01 M DOSO and ≥81 in the presence of 0.03 M PetrSO [139]. The separation factors in the absence of a synergist are not given. 2. Effect of Extraction Variables Available data allow an assessment of the effect of the sulfoxide structure. In the extraction of U(VI) by 4-benzoyl-2,4-dihydro-5-methyl-2-phenyl-3H-pyrazol-3thione the synergistic action in toluene decreases in the order PetrSO>>DOSO>TBP [133]. In the extraction of Nd(III) by 3,5-di(tert-butyl)salicylic acid, the synergistic action decreases in the order dicyclohexyl sulfoxide~cyclohexyl hexyl sulfoxide~ cyclohexyl octyl sulfoxide~cyclooctyl octyl sulfoxide>DBSO~DHxSO>DOSO >cyclopentyl octyl sulfoxide>DBzSO>DEHSO~DΦSO. Similar orders have been found with 5-hexyl- and 3,5-diisopropylsalicylic acid as the acidic component [140]. With 0.1 M 4-benzoyl-2,4-dihydro-5-methyl-2-phenyl-3H-pyrazol-3-thione in toluene the separation factor αU(VI)/Th(IV) is 760 in the presence of 0.03 M PetrSO and 480 in the presence of 0.01 M DOSO [133]. Also comparison of sulfoxides with phosphoryl extractants is possible. In the extraction of Eu(III) by 1-phenyl-3-methyl-4-(2-chlorobenzoyl)-5-pyrazolone in xylene the synergistic action increases in the order TBPNH4ClO4, and addition of one of these salt up to a concentration of 2 M enhances the DU(VI) value by a factor of 3–5. The phenomenon is ascribed to suppression of the water activity in the system by the salts [148]. 3. Nature of Extracted Complexes and Thermodynamic Functions Besides slope analysis, also dependencies of the type shown in Fig. 21 can be used for the elucidation of the composition of the extracted synergistic complexes. The position of the maximum on the curves can show the ratio of the acidic and sulfoxide extractant molecules participating in the formation of the synergistic complex.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

229

Examples of slope analysis in Fig. 22 shows that, in the measured range of the synergist concentration, a synergistic complex can predominate in the organic phase or it can be present together with a simple nonsynergistic complex. It is quite evident in Fig. 22 that synergistic complexes ThA 4 · B and NpA 4 · B are predominantly extracted, while the complex PuA4 · B coexists with PuA4. The extracted complexes have typically a simple composition in the extraction of actinides(IV, VI), where no anions of the aqueous medium participate in the formation of the extracted complex. The complexes contain one sulfoxide molecule and are • • • • •

UO2A2 · B with HA=TTA and B=DBzSO or DΦSO in benzene from 0.01 M HCl[130] UO2A2 · B with HA=1-phenyl-3-methyl-4-acyl-5-pyrazolone (acyl=acetyl, trifluoroacetyl, benzoyl, 2-chlorobenzoyl, or 4-nitrobenzoyl) and B=DOSO in chloroform from 0.5 M HCl [131] UO2A2 · B with HA=1-phenyl-3-methyl-4-benzoyl-5-pyrazolone and B= DΦSO in benzene from 0.05 M HNO3 [132] UO2A2 · B, with HA=4-benzoyl-2,4-dihydro-5-methyl-2-phenyl-3H-pyrazol3-thione and B=PetrSO in toluene at pH 2.04 [133] ThA4 · B, NpA4 · B, and PuA4 · B with HA=TTA and B=DOSO in benzene from 0.1 M HClO4 [136]

Complexes of Zr(IV) and Hf(IV) typically contain anions of the aqueous medium and one to three sulfoxide molecules. They can be written as • • • • •

ZrA2Cl2 · 2B and ZrA2Cl(ClO4) · 2B with HA=TTA and B=DPSO in CCl4 from unclearly specified aqueous solutions [104] ZrA2(OH)2 · B with HA=TTA and B=DPSO in benzene from 2 M HCl+2 M NaCl [91] HfA(NO3)3 · B with HA=TTA and B=DPSO in CCl4 from 9 M HNO3 [82] HfA3(ClO4) · B and HfA2(ClO4)2 · B with HA=TTA and B=DOSO in CCl4 from 6 M HClO4 [82] HfACl3 · 3B, HfA2Cl2 · 2B, and HfA3Cl · B with HA=TTA and B=DPSO in CCl4 from 7 M HCl [103]

Complexes of lanthanides(III), Am(III), and Zn(II) can contain complexing inorganic anions like thiocyanates, together with up to three sulfoxide molecules. They do not contain noncomplexing anions of the aqueous medium, such as perchlorates, and then the number of sulfoxide molecules is one to two. Reported compositions are • •

AmA3 · B, AmA2(SCN) · 2B, and AmA(SCN)2 · 3B with HA=TTA and B= DPSO in benzene from 1 M NaSCN at pH 3 [123] AmA2(NO3) · 2B with HA=4-benzoyl-2,4-dihydro-5-methyl-2-phenyl-3Hpyrazol-3-thione and B=DOSO or PetrSO in toluene from 0.1 M NaNO3 at pH 3.6 [139]

Copyright © 2004 by Marcel Dekker, Inc.

230

Kolarik

Table 20 Equilibrium Constants of the Formation of Synergistic Complexes in the Organic Phasea

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

231

Table 20 Continued

a

Ks1=[UO2A2 · B][UO2A2]-1[B]-1 and KS2=[UO2A2 · 2B][UO2A2]-1 [B]-2 (italics).

• • • • • • • • •

MA3 · B and MA3 · 2B with M=Am and Eu, HA=TTA, and B=DBSO in CCl4 from 1 M (H,Na)ClO4 [4] MA3 · B and MA3 · 2B with M=Nd and Eu, HA=4,4,4-trifluoro-1-phenyl1,3-butanedione and B=DEHSO in benzene from 0.1 M thiocyanate at pH 3 [145] MA3 · B and MA3 · 2B with M=Nd, Eu, and Lu, HA=1-phenyl-3-methyl-4trifluoroacetyl-5-pyrazolone from 0.01 M chloroacetate buffer at pH 2.7 [146] NdA3 · B and NdA3 · 2B with HA=3-phenyl-4-benzoyl-5-isoxazolone and B =DOSO, DEHSO, or DΦSO in xylene from 0.1 M NaClO4 at pH 3 [147] EuA2Cl · B with HA=1-phenyl-3-methyl-4-(2-cHorobenzoyl)-5-pyrazolone and B=DOSO or dicyclohexyl sulfoxide in xylene from 0.1 M (HCl+NaCl) at pH 1.5 [141] MA3 · B with M=Tb and Tm, HA=3-phenyl-4-benzoyl-5-isoxazolone, and B=DOSO, DEHSO, or DΦSO in xylene from 0.1 M NaClO4 at pH 3 [147] TmA3 · B and TmA3 · 2B with HA=TTA and B=DBSO in benzene from 1 M NaClO4 at pH3 [119] TmA3 · B, TmA3 · 2B, TmA2(SCN) · 2B, and TmA (SCN)2 · 3B with HA=TTA and B=DBSO in benzene from 1 M NaSCN at pH 3 [119] MA3 · B with M=Sc and Lu, HA=TTA, and B=DBSO in CCl4 from 1 M (H,Na)ClO4 [4]

Copyright © 2004 by Marcel Dekker, Inc.

232 • •

Kolarik LuA3 · B with HA=4,4,4-trifluoro-1-phenyl-1,3-butanedione and B=DEHSO in benzene from 0.1 M thiocyanate at pH 3 [145] ZnA2 · B with HA=1-phenyl-3-methyl-4-benzoyl-5-pyrazolone and B= DEHSO in benzene from 0.01 M acetate buffer at pH 3.75 [114]

An unusual complex is reported to be formed by the ion with di(2-ethylhexyl)phosphoric acid (HA) and PetrSO in benzene. Its composition is VO2(HA2) (HA)2 · B, where the monoionized dimer forms the usual 8-membered ring and neutralizes the charge of the ion, the two monomeric molecules HA are bound to the ion via the phosphoryl O atoms, and the PetrSO molecule is bonded to one of the HA molecules via a hydrogen bond of its OH group [138]. The extent of the synergistic action of a sulfoxide is characterized by the formation constants of the synergistic complexes in the organic phase. Data are available only for mono- and disolvate complexes, where the constants are KS1= [UO2A2 · B][UO2A2]-l[B]-1 and KS2=[UO2A2 . 2B][UO2A2]-1[B]-2. The synergistic distribution ratio is then DS=Da(1+KS1[B]+Ks2[B]2) with Da being the distribution ratio in the absence of the synergist. Published formation constants are gathered in Table 20. It is seen there that the stability of the complex UO2A2 · B depends not only on the nature of the synergist but also on the anion A- of the acidic component. As can be expected, infrared spectra show that the sulfoxide molecule is bound to the extracted metal ion via the oxygen atom. The S→O stretch frequency in the TTA complexes UO2A2 · 2B is shifted to lower values, by 55 cm-1 with B=DΦSO and by 58 cm-1 with B=DBzSO. In spite of the small difference between the shifts, it is concluded that DBzSO is bound more strongly to the uranyl ion than DΦSO [130]. Thermodynamic functions of the complex formation of U(VI) and Pu(VI) are given in Table 21.

Table 21 Thermodynamic Functions of the Reaction MO2A2+B=MO2A2 · B in the Organic Phase and of the Biphasic Reaction (italics)a

a

Benzene diluent.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

233

NOMENCLATURE A. Symbols aw, aHX, B DM, DB SE Kex LOC M αM/M’

activity of water, an acid HX, and X- ions, respectively extractant molecule in the formula of a solvate distribution ratio of a metal (the valence state can also be characterized by the subscript) or of a sulfoxide synergistic enhancement, DB1B2(DB1+DB2)-1, with DB1, DB2 and DB1B2 denoting the DM value with extractant 1, extractant 2, and a mixture of the extractants equilibrium constant of an extraction reaction on the molarity scale limiting organic concentration of the metal metal, eventually with valence specified in parentheses separation factor for metals M and M’ defined as DM/DM’

B. Abbreviations of Sulfoxide Extractants DBSO DBzSO DDSO DHpSO DHxSO DiPSO DLSO DNSO DOSO DPSO DΦSO PetrSO

dibutyl sulfoxide dibenzyl sulfoxide didecyl sulfoxide diheptyl sulfoxide dihexyl sulfoxide diisopentyl sulfoxide didodecyl sulfoxide dinonyl sulfoxide dioctyl sulfoxide dipentyl sulfoxide diphenyl sulfoxide petroleum sulfoxides

C. Abbreviations of Other Extractants and a Diluent MiBK PMBP TBP TCE TTA

2-methyl-4-pentanone, frequently called methyl isobutyl ketone 1-phenyl-3-methyl-4-benzoyl-5-pyrazolone tributyl phosphate 1,1,2-trichloroethane 1-(2-thienyl)-4,4,4-trifluoro-1,3-butanedione, frequently called thenoyltrifluoroacetone

Copyright © 2004 by Marcel Dekker, Inc.

234

Kolarik

APPENDIX: PHYSICAL PROPERTIES OF SULFOXIDES A. Melting and Boiling Points

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

a

Distillation from a reaction mixture.

B. Densities, Dipole Moments, and Refractive Indices

Copyright © 2004 by Marcel Dekker, Inc.

235

236

a b

Kolarik

In benzene. At 24.3°C

c

C. Physical Properties of a Solution 1.0 M DHpSO in TCE has at 20 °C a density of 1.280 g cm-3 and a viscosity of 1.775 cp, and its interfacial tension with water is 17.5 dynes cm-1 [34].

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

Korpak, W. Nukleonika 1962, 7, 715–723. Korpak, W. Nukleonika 1964, 9, 1–10. Korpak, W. Nukleonika 1963, 8, 747–754. Sekine, T.; Dyrssen, D.J. Inorg. Nucl. Chem. 1967, 29, 1481–1487. Nikitin, Yu.E.; Murinov, Yu.I.; Rozen, A.M. Usp. Khim. 1976, 45, 2233–2252. Laurence, G.C.R.Acad. Sci. Paris Ser. C 1969, 269, 352–354. Chengye, Yuan; Haiyan, Long; Enxin, Ma; Wuhua, Chen; Xiaomin, Yan. Scient. Sinica 1984, 27, 887–897. McDowell, W.J.; Harmon, H.D. J.Inorg. Nucl. Chem. 1971, 33, 3107–3117. Siddal, T.H. III; Brown, D.A. Science and Technology of Tributyl Phosphate. In Synthesis, Properties, Reactions and Analysis, Schulz, W.W., Navratil, J.D., Talbot, A.E., Eds.; CRC Press: Boca Raton, FL, 1984; Vol. 1, 15–24. Sharipov, A.Kh. Khim. Prom. 2000; 24–32. Shen, Chaohong; Bao, Borong; Bao, Yizhi; Wang, Gaodong; Qian, Ju; Cao, Zhengbei. J. Radioanal. Nucl. Chem. 1994, 178, 91–98. Mikhailichenko, A.I.; Sokolova, N.P.; Sulaimankulova, S.K.; Teterin, E.G. Radiokhimiya 1973, 15, 693–697; Soviet Radiochem. 1973, 15, 699–703. Zhang, Zuokui; Chen, Zhengkang. He Huaxue Yu Fangshe Huaxue 1983, 5, 309–314.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

237

14. Shanker, R.; Venkateswarlu, K.S.J. Inorg. Nucl. Chem. 1970, 32, 229–237. 15. Rozen, A.M.; Murinov, Yu.I.; Nikitin, Yu.E.; Pilyugin, V.S. Radiokhimiya 1974, 16, 118-120; Soviet Radiochem. 1974, 16, 115–117. 16. Preston, J.S.; du Preez, A.C. Solvent Extr. Ion Exch. 2001, 19, 581–595. 17. Moyer, B.A.; Caley, C.E.; Baes, C.F. Jr. Solvent Extr. Ion Exch. 1988, 6, 785–817. 18. Moyer, B.A.; Baes, C.F. Jr.; Case, F.I.; Driver, J.L. Solvent Extr. Ion Exch. 2001, 19, 757–790. 19. Marcus, Y.; Kolarik, Z.J. Chem. Eng. Data 1973, 18, 155–163. 20. Sato, T.; Ishikawa, I.; Sato, K. Shigen to Sozai 1998, 114, 33–39. 21. Moyer, B.A.; McDowell, W.J.; Baes, C.F. Jr.; Case, G.N.; Case, F.I. Solvent Extr. Ion Exch. 1991, 9, 833–864. 22. Laurence, G.J.; Chaieb, M.T.; Talbot, J. Proceedings of the International Solvent Extraction Conference ISEC 71, The Hague, 1971; 1150–1160. 23. Mikhailov, V.A. Proceedings of the International Solvent Extraction Conference ISEC ’77, Toronto, 1977; 52-60. 24. Chaieb, M.T.; Laurence, G. Energ. Nucl. (Paris) 1971, 13, 106–113. 25. Shukla, J.P.; Pai, S.A.; Subramanian, M.S. J. Radioanal. Chem. 1980, 56, 53–63. 26. Pai, S.A.; Shukla, J.P.; Khopkar, P.K.; Subramanian, M.S. J. Radioanal. Chem. 1978, 42, 323–334. 27. Rozen, A.M.; Murinov, Yu.I.; Nikitin, Yu.E.; Mineeva, N.Z. Radiokhimiya 1972, 14, 750–752; Soviet Radiochem. 1972, 14, 771-772. 28. Bo-Rong, Bao; Ya-Zhi, Bao; Chao-Hong, Shen; Gao-Dong, Wang; Jun, Qian; ZhengBai, Cao J. Radioanal. Nucl. Chem. 1992, 162, 391–398. 29. Shukla, J.P.; Nagar, M.S.; Subramanian, M.S. Proceedings of the International Solvent Extraction Conference ISEC ‘83, Denver, 1983; 460–461. 30. Rozen, A.M.; Murinov, YuI; Nikitin, YuE. Radiokhimiya 1970, 12, 355–361; Soviet Radiochem. 1970, 12, 322-326. 31. Laurence, G.J. Inorg. Nucl. Chem. 1970, 32, 3065–3071. 32. Rozen, A.M.; Murinov, YuI; Nikitin, YuE. Radiokhimiya 1970, 12, 513–516; Soviet Radiochem. 1970, 12, 477-479. 33. Preston, J.S.; du Preez, A.C. J. Chem. Tech. Biotechnol. 1997, 69, 86–92. 34. Laurence, G.; Chaieb, M.T.; Habsieger, D.; Michel, P.; Talbot, J. Proceedings of the International Solvent Extraction Conference ISEC ‘74, Lyon, 1974; 2271–2289. 35. Laurence, G.; Chaieb, M.T. Bull. Soc. Chim. Fr. 1970; 2481–2485. 36. Zhu, Guohui. Yuanzineng Kexue Jishu 1987, 21, 560–564. 37. Yanzhao, Yang; Yonghui, Yang; Sixiu, Sun; Borong, Bao. Shandong Daxue Xuebao, Ziran Kexueban 1998, 33, 309–313. 38. Shukla, J.P.; Pai, S.A.; Subramanian, M.S. Sep. Sci. Technol. 1979, 14, 883–894. 39. Moyer, B.; McDowell, W.J.; Case, G.N. Proceedings of the International Solvent Extraction Conference ISEC ‘83, Denver, 1983; 441-442. 40. Wang, Hanzhang; Qian, Binyan; Zong, Yaping; Gu, Jiansheng. He Huaxue Yu Fangshe Huaxue 1987, 9, 43–45. 41. Yang, Yan-Zhao; Sun, Si-Xiu; Bao, Bo-Rong. Nucl. Sci. Techn. 1998, 9, 243–245. 42. Shukla, J.P.; Kedari, C.S. J. Radioanal. Nucl. Chem. 1996, 207, 93–105. 43. Yang, Yan-Zhao; Yang, Yong-Hui; Jiang, Xu-Chuan; Sun, Si-Xiu; Bao, Bo-Rong J. Radioanal. Nucl. Chem. 1999, 240, 335–338. 44. Yang, Yan-Zhao; Sun, Si-Xiu; Yang, Yong-Hui; Zhang, Lei; Yu, Shao-Ning; Bao, BoRong J. Radioanal. Nucl. Chem. 1999, 241, 231–234.

Copyright © 2004 by Marcel Dekker, Inc.

238

Kolarik

45. Wang, Jinhua; Zhou, Zuming; Xu, Dimin; Chen, Yude; Mao, Jiajun. He Huaxue Yu Fangshe Huaxue 1992, 14, 226–231. 46. Yu, Shao-Ning; Fan, Zheng-Ping; Ma, Li; Yang, Yan-Zhao; Bao, Bo-Rong J. Radioanal. Nucl. Chem. 1998, 232, 249–251. 47. Yu, Shao-Ning; Fan, Zheng-Ping; Ma, Li; Yang, Yan-Zhao; Bao, Bo-Rong. Solvent Extraction for the 21st Century. Proceedings of ISEC ‘99, Barcelona, 1999; 353-357. 48. Yang, Yan-Zhao; Yong-Hui, Yang; Cao, Xiao-Shi; Sun, Si-Xiu. Radiochim. Acta 2000, 88, 355–359. 49. Yang, Yanzhao; Yang, Yonghui; Sun, Sixiu; Yu, Shaoning; Han, Jingtian; Bao, Borong. Hejishu 1999, 22, 187–190. 50. Shanker, R.; Venkateswarlu, K.S.J. Inorg. Nucl. Chem. 1970, 32, 2369–2381. 51. Wang, Hanzhang; Pan, Jianhua; Gu, Jiansheng. He Huaxue Yu Fangshe Huaxue 1991, 13, 157–162. 52. Zhou, Zuming; Ji, Jianguo; Chen, Guoming; Du, Huifang; Mao, Jiajun. He Huaxue Yu Fangshe Huaxue 1994, 16, 13–17. 53. Horng, Jiin-Shiung; Maa, Jer-Ru. Hydrometallurgy 1984, 12, 355–364. 54. Petkovic, D.M.; Kopecni, M.M.; Janicijevic, P.M. J. Serb. Chem. Soc. 1991, 56, 479–488. 55. Li, Yi-Gui; Teng, Teng; Cheng, Zhen-Hua; Zhang, Liang-Ping. Hydrometallurgy 1982, 8, 273–287. 56. Yang, Yan-Zhao; Sun, Si-Xiu; Feng, Sheng-Yu. J. Radioanal. Nucl. Chem. 2002, 251, 503–506. 57. Rozen, A.M.; Murinov, Yu.I.; Nikitin, Yu.E.; Krivinogov, V.P. Radiokhimiya 1971, 13, 891–893; Soviet Radiochem. 1971, 13, 917–918. 58. Yang, Yan-Zhao; Sun, Si-Xiu; Zhang, Lei; Yang, Yong-Hui; Bao, Bo-Rong. J. Radioanal. Nucl. Chem. 1999, 240, 383–385. 59. Han, Jing-Tian; Sun, Guo-Xin; Shao, Hua; Chen, Shu-Feng; Bao, Bo-Rong. J. Radioanal. Nucl. Chem. 1999, 242, 821–824. 60. Subramanian, M.S. J. Radioanal. Nucl. Chem. 1994, 187, 91–97. 61. Yang, Yanzhao; Yang, Yonghui; Yan, Yong; Sun, Sixiu; Bao, Borong. Shandong Daxue Xuebao, Ziran Kexueban 1998, 33, 76–80. 62. Shukla, J.P.; Kedari, C.S. Radiochim. Acta 1994, 64, 179–184. 63. Xia, Y.X.; Bao, B.R.; Yang, J.F. J. Radioanal. Nucl. Chem. 1988, 125, 165–173. 64. Yang, Yan-Zhao; Feng, Sheng-Yu. J. Radioanal. Nucl. Chem. 2002, 253, 161–164. 65. Yang, Yan-Zhao; Tian, Sheng-Yun; Sun, Si-Xiu; Feng, Sheng-Yu. J. Radioanal. Nucl. Chem. 2002, 252, 171–173. 66. Yi, M in; Chen, Dingfong; Wang, Wenqing. He Huaxue Yu Fangshe Huaxue 1982, 4, 73–80. 67. Dong, Yan; Yang, Yan-Zhao; Fan, Wei-Liu; Sun, Si-Xiu. Nucl. Sci. Tech. 2001, 12, 270–275. 68. Yang, Yan-Zhao; Feng, Sheng-Yu. J. Radioanal. Nucl. Chem. 2002, 253, 149–151. 69. Comor, J.J.; Kopecni, M.M.; Petkovic, D.M. J. Serb. Chem. Soc. 1995, 60, 1035– 1041. 70. Torgov, V.G; Nikolaev, A.V.; Mikhailov, V.A.; Stadnikova, L.G.; Kotlyarevskii, I.L..Zh. Neorg. Khim. 1968, 13, 228–231; Russ. J. Inorg. Chem. 1968, 13, 119–121. 71. Mahajan, G.R.; Prabhu, D.R.; Shukla, J.P.; Nair, G.M. Thermochim. Acta 1992, 196, 357–365.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

239

72. Pai, S.A.; Shukla, J.P.; Subramanian, M.S. J. Inorg. Nucl. Chem. 1981, 43, 1045– 1047. 73. Mahajan, G.R.; Murali, M.S.; Prabhu, D.R.; Shukla, J.P.; Nair, G.M.; Natarajan, P.R. Thermochim. Acta 1991, 180, 185–190. 74. Khan, M.H.; Hasany, S.M.; Ali, A.; Khan, M.A. J. Radioanal. Nucl. Chem. 1995, 198, 409–421. 75. Rozen, A.M.; Murinov, Yu.I.; Nikitin, Yu.E. Radiokhimiya 1970, 12, 516–518; Soviet Radiochem. 1970, 12, 480-482. 76. Mikhailov, V.A.; Torgov, V.G.; Gil’bert, E.N.; Mazalov, L.N.; Nikolaev, A.V. Proceedings of the International Solvent Extraction Conference ISEC ‘71. The Hague, 1971; 1112-1119. 77. Prabhu, D.R.; Mahajan, G.R.; Murali, M.S.; Shukla, J.P.; Nair, G.M.; Natarajan, P.R. J. Radioanal. Nucl. Chem. 1992, 162, 91–97. 78. Shukla, J.P.; Kedari, C.S. Radiochim. Acta 1992, 56, 21–24. 79. Khan, M.H.; Hasany, S.M.; Khan, M.A.; Ali, A. Radiochim. Acta 1994, 65, 239–243. 80. Kulesza, A. Chem. Anal. (Warsaw) 1970, 15, 577–583. 81. Khan, M.H.; Ali, A. Radiochim. Acta 2002, 90, 297–301. 82. Hála, J.; Rosický, L. Solvent Extr. Ion Exch. 1989, 7, 87–103. 83. Shukla, J.P; Kedari, C.S. J. Radioanal. Nucl. Chem. 1992, 157, 355–366. 84. Mikhailichenko, A.I.; Kotliarov, R.V.; Sokolova, N.P.; Abramov, L.A. Proceedings of the International Solvent Extraction Conference ISEC ’74, Lyon, 1974; 1059–1072. 85. Shukla, J.P.; Singh, R.K.; Sawant, S.R.; Varadarajan, N. Anal. Chim. Acta 1993, 276, 181–187. 86. Shukla, J.P.; Nagar, M.S.; Subramanian, M.S. Sep. Sci. Technol. 1982-83, 17, 1577– 1585. 87. Mikhailov, V.A.; Obraztsova, I.I. In Khimiya Ekstraktsii; Mikhailov, V.A.; Ed.; Nauka Sibirskoe Otdelenie: Novosibirsk, 1984; 186–193. 88. Chakravortty, V.; Dash, K.C.; Mohanty, S.R. Radiochim. Acta 1986, 40, 89–94. 89. Reddy, A.S.; Reddy, L.K. Sep. Sci. 1977, 12, 641–644. 90. Brodskaya, G.A.; Gureev, E.S.; Usachenko, V.S.; Zolotov, YuA. Zh. Anal. Khim. 1974, 29, 2209–2215;]. Anal. Chem. USSR 1974, 29, 1900–1905. 91. Panda, C.R.; Chakrawortty, V.; Dash, K.C.; Mohanty, S.R. Solvent Extr. Ion Exch. 1986,4, 969–985. 92. Reddy, A.S.; Ramakrishna, V.V.; Patil, S.K. Radiochem. Radioanal. Lett. 1977, 28, 445–452. 93. Chakravortty, V.; Mohanty, S.R. Radiochem. Radioanal. Lett. 1979, 41, 275–278. 94. Mohanty, S.R.; Reddy, A.S. J. Inorg. Nucl. Chem. 1975, 37, 1791–1795. 95. Gaur, P.K.; Mohanty, S.R. Naturwiss. 1963, 50, 614. 96. Kennedy, D.C.; Fritz, J.S. Talanta 1970, 17, 823–835. 97. Chakrabortty, M.; Chakravortty, V.; Mohanty, S.R. J. Radioanal. Nucl. Chem. 1979, 53, 131–143. 98. Mishra, P.K.; Chakravortty, V.; Dash, K.C.; Das, N.R.; Bhattacharyya, S.N. J. Radioanal. Nucl. Chem. 1989, 131, 281–287. 99. Zhang, Banglao. Shaanxi Shifan Daxue Xuebao, Ziran Kexueban 1998, 26, 70–73. 100. Preston, J.S.; du Preez, A.C. Solvent Extr. Ion Exch. 2002, 20, 359–374. 101. Ji, Jianguo; Zhou, Zuming; Mao, Jiajun. Guijinshu 1996, 17, 1–5. 102. Mohanty, S.R.; Reddy, A.S. J. Inorg. Nucl. Chem. 1975, 37, 1977–1982. 103. Reddy, A.S.; Reddy, B.R. J. Inorg. Nucl. Chem. 1981, 43, 2570–2572.

Copyright © 2004 by Marcel Dekker, Inc.

240

Kolarik

104. Reddy, A.S.; Reddy, B.R. J. Radioanal. Chem. 1978, 47, 77–82. 105. Cheng, Fei; Gu, Guobang; Zhang, Zhenmin; Long, Tiwu; Pang, Shuanggen. Huanan Ligong Daxue Xuebao, Ziran Kexueban 1995, 23, 98–103. 106. Sato, T.; Ishikawa, I.; Sato, K. Shigen to Sozai 2000, 116, 44–48. 107. Francis, T.; Ramamohan, T.; Reddy, M.L. J. Chem. Technol. Biotechnol. 2001, 76, 737–742. 108. Zhang, Kecheng; Chen, Jing. Guijinshu 1995, 16, 6–13. 109. Qian, Hui-Fen; Zhai, Li-Xin; Liu, Jian-Lan. Nanjing Huagong Daxue Xuebao, Ziran Kexueban 2001, 23, 54–57. 110. Qian, Huifen; Liu, Jianlan. Huagong Shikan 2001, 15, 33–36. 111. Reddy, A.S.; Reddy, L.K. Sep. Sci. Technol. 1980, 15, 1263–1269. 112. Reddy, A.S.; Reddy, L.K. J. Inorg. Nucl. Chem. 1977, 39, 1683–1687. 113. Reddy, A.S.; Reddy, L.K. Radiochem. Radioanal. Lett. 1977, 28, 347–353. 114. Reddy, M.L.P.; Preetha, G.P.; Meena, C.V.; Prasada Rao, T.; Iyer, C.S.P.; Damodaran, A.D. J. Radioanal. Nucl. Chem. 1996, 211, 305–316. 115. Behera, P.; Mishra, R.; Chakrawortty, V.J. Radioanal. Nucl. Chem. 1993, 173, 161– 169. 116. Fritz, J.S.; Kennedy, D.C. Talanta 1970, 17, 837–843. 117. Manchanda, V.K.; Mohapatra, P.K. Radiochim. Acta 1995, 69, 81–85. 118. Rozen, A.M.; Murinov, Yu.I.; Nikitin, Yu.E.; Abramova, A.A. Radiokhimiya 1972, 14, 752–754; Soviet Radiochem. 1972, 14, 773-774. 119. Ramakrishna, V.V.; Patil, S.K.; Reddy, L.K.; Reddy, A.S. J. Radioanal. Chem. 1978, 47, 57–76. 120. Reddy, L.K.; Reddy, A.S.; Reddy, M.L.P.; Ramamohan, T.R.; Damodaran, A.D. J. Radioanal. Nucl. Chem. 1993, 170, 89–97. 121. Li, Huanran; Wu, Qingzhu; Rong, Qingxin. Zhongshan Daxue Xuebao, Ziran Kexueban 1997, 36, 77–82. 122. Li, Huanran; Zhu, Xiaowen; Wu, Jinghong; Xu, Wanyi; Lin, Yaobing. Fenxi Huaxue 1999, 27, 708–711. 123. Ramakrishna, V.V.; Patil, S.K.; Reddy, L.K.; Reddy, A.S. J. Inorg. Nucl. Chem. 1979, 41, 108–113. 124. Santhi, B.P.; Reddy, M.L.P.; Ramamohan, T.R.; Damodaran, A.D. Radiochim. Acta 1994, 64, 121–125. 125. Reddy, M.L.P.; Sujatha, S.; Ramamohan, T.R.; Rao, T.P.; Iyer, C.S.P.; Damodaran, A.D. Radiochim. Acta 1995, 69, 201–204. 126. Mohapatra, P.K.; Manchanda, V.K. J. Radioanal. Nucl. Chem. 1999, 240, 159–164. 127. Kumar, S.; Koganti, S.B. Solvent Extr. Ion Exch. 1998, 16, 829–841. 128. Pai, S.A.; Shukla, J.P.; Subramanian, M.S. J. Radioanal. Chem. 1982, 74, 31–38. 129. Shukla, J.P.; Kedari, C.S. Nucl. Sci. J. 1995, 32, 92–100. 130. Subramanian, M.S.; Viswanatha, A.J. Inorg. Nucl. Chem. 1969, 31, 2575–2585. 131. Mao, Jia-Jun; Jia, Wei-Jie; Shao, Xing-Zhi; Chen, Yu-De J. Radioanal. Nucl. Chem. 1991, 147, 287–295. 132. Pai, S.A.; Subramanian, M.S. J. Radioanal. Nucl. Chem. 1985, 89, 423–433. 133. Yu, Shaoning; Li, Ma; Bao, Borong. J. Radioanal. Nucl. Chem. 1999, 241, 347–349. 134. Lohithakshan, K.V.; Mithapara, P.D.; Pai, S.A.; Aggarwal, S.K. Radiochim. Acta 1996, 75, 211–217. 135. Lohithakshan, K.V.; Mithapara, P.D.; Pai, S.A.; Aggarwal, S.K.; Jain, H.C. Radiochim. Acta 1996, 73, 185–190.

Copyright © 2004 by Marcel Dekker, Inc.

Sulfoxide Extractants and Synergists

241

136. Sajun, M.S.; Ramakrishna, V.V.; Patil, S.K. J. Radioanal. Chem. 1981, 63, 57–63. 137. Mansingh, P.S.; Chakrawortty, V.; Dash, K.C. Radiochim. Acta 1996, 73, 139–143. 138. Gu, Guobang; Cheng, Fei; Yang, Xinrong; Zhang, Zhenmin; Long, Tiwu. Huanan Ligong Daxue Xuebao, Ziran Kexueban 1997, 25, 85–91. 139. Yu, Shao-Ning. Nucl. Sci. Techn. 2000, 11, 192–196. 140. Preston, J.S.; du Preez, A.C. Solvent Extr. Ion Exch. 1996, 14, 755–772. 141. Zhou, Meicun; Sun, Xiaoyu; Mao, Jiajun. He Huaxue Yu Fangshe Huaxue 1997, 19, 5–10. 142. Mohanty, I.; Murlidhar, J.; Chakrawortty, V.J. Radioanal. Nucl. Chem. 1998, 227, 111–116. 143. Zhang, Anyun. Shaanxi Shifan Daxue Xuebao, Ziran Kexueban 1999, 27, 51–54, 61. 144. Zhang, Anyun. Solvent Extr. Ion Exch. 2000, 18, 1189–1197. 145. Reddy, M.L.P.; Sujatha, S.; Varma, R.L.; Ramamohan, T.R.; Rao, T.P.; Iyer, C.S.P.; Damodaran, A.D. Talanta 1997, 44, 97–103. 146. Sujatha, S.; Reddy, M.L.P.; Varma, R.L.; Ramamohan, T.R.; Rao, T.P.; Iyer, C.S.P.; Damodaran, A.D. J. Chem. Eng. Japan 1996, 29, 187–190. 147. Sahu, S.K.; Chakrawortty, V.; Reddy, M.L.P.; Ramamohan, T.R. Radiochim. Acta 1999, 85, 107–111. 148. Subramanian, M.S.; Pai, S.A. J. Inorg. Nucl. Chem. 1970, 32, 3677–3685. 149. Venier, C.G.; Squires, T.G.; Chen, Y.; Hussmann, G.P.; Shei, J.C.; Smith, B.F. J. Org. Chem. 1982, 47, 3773–3774. 150. Weibull, B. Arkiv. Kemi. 1958, 3, 171–223. 151. Dictionary of Organic Compounds. 5th Ed.; 2nd suppl.; Chapman and Hall: New York, 1984; 185 pp.

Copyright © 2004 by Marcel Dekker, Inc.

5 Extraction with Metal Bis(dicarbollide) Anions

Metal Bis(dicarbollide) Extractants and Their Applications in Separation Chemistry Jirí Rais Nuclear Research Institute Rez plc, Rez, Czech Republic Bohumír Grüner Institute of Inorganic Chemistry, Czech A.cademy of Sciences, Rez, Czech Republic

I.

INTRODUCTION

A.

History of Metal Bis(dicarbollide) Extractants Research

This review resulted from rather broad studies of two Czech institutes, namely Institute of Inorganic Chemistry, Czech Academy of Sciences, Rez, IIC (B.G.) and Nuclear Research Institute, Rez, NRI (J.R.). The collaboration of the two institutes started not only because of the same location at Rez near Prague, Czech Republic, but mainly from the common interests in development of effective extractants for their use in the area of nuclear applications. Although the two areas of the review— synthesis and properties of metal bis(dicarbollide) anions (IIC) and extraction research using these anions (NRI)—differ in their subjects, techniques and even in style of narration, the two must be put together in order to have a complete picture of the subject. Section II, Tables 1–11, Figs. 1 and 2 and the reaction schemes in the review were written and drawn by B.G. The early studies with various hydrophobic anions have their fundamental standing, but the discovery of the excellent extraction properties of metal 243 Copyright © 2004 by Marcel Dekker, Inc.

244

Rais and Grüner

bis(dicarbollide) anions marked a qualitatively new step in the research. The first stable metal bis(dicarbollide)s were synthesized and described by Hawthorne in the United States already in 1965 [1], but their use as perfect hydrophobic anions waited until 1976 when their extraction properties were first published [2]. (For the sake of completeness, Schlaikjer already in 1965 studied the extraction of Cs+ and Ba2+ with polyhedral and anions [3].) The cobalt bis(dicarbollide) anion [(1,2-C2B9H11)2-3-Co]- (1) was very soon established as an ideal hydrophobic anion for extractions by the ion pair mechanism [2]. Its advantages are (1) very high hydrophobicity, (2) extreme acidity of the derived acid H+1- that behaves as a superacid and is fully dissociated even in media of ε =10 to 15, (3) high chemical and radiation stability. This is so especially for the chloroprotected cobalt bis(dicarbollide) derivative. The cobalt bis(dicarbollide) anion surpasses by an order of magnitude the until now examined anions usable for the extraction of radioactive cesium and soon after the publication [2] the process of Cs isolation with this ion was patented by Czech scientists [4]. The important supporting finding of polyethylene glycols as strong synergists for simultaneous extraction of strontium with cobalt bis(dicarbollide)s [5, 6] enabled the construction of a conceptual process for the isolation of both cesium and strontium from highly acidic aqueous solutions. Besides the patenting in the former Czechoslovakia, the method of isolation of 137 Cs and 90Sr was patented by the Czech scientists in those early commencements of the research also in the former USSR [7] and soon a long-standing cooperation between the Radium Institute of St. Petersburg (the former Leningrad) and NRI, Rez started. In this cooperation, the chemistry of the developed process was studied in Rez whereas the hot cell experiments were executed in the hot cell facility at Gatchina, Russia. A number of universities from the Czech side joined the program, devoting themselves to the tasks of studying the radiation stabilities of proposed organic mixtures, explosive hazards, extraction of other valuable components, physicochemical parameters of the extractants in order to check their composition, etc. The preparation of the extractant for the planned plant tests [200 kg of Cs salt of chloroprotected cobalt bis(dicarbollide)] and tests of purity were also performed in the former Czechoslovakia. The process was developed in 1984 to the stage enabling its pilot plant testing and this was performed in a 6-month run at the reprocessing plant Mayak in Russia. The process was repatented with Russian scientists [8] and this stage of development was thus successfully accomplished. The cold war with its complete lack of information flows between the two world blocks caused that a new initiative in the United States for studies of cobalt bis(dicarbollide)s appeared only at the Los Alamos National Laboratory (LANL) in 1990 [9]. However, the LANL was not involved in the main stream of the development of cobalt bis(dicarbollide) technology, which on the contrary has continued for many years between the U.S. Department of Energy (DOE), the Idaho National Engineering and Environmental Laboratory (INEEL), and the Radium Institute of St. Petersburg (RI). The research teams formed two leading working groups: one group is that of DOE, INEEL, and RI and the other one is that

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

245

of IIC, NRI, and other European institutes and universities in the frame of Euro projects. It is characteristic for the, in some aspects surreptitious, collaboration between the United States (DOE) and Russia (RI) that inventor of cobalt bis(dicarbollide)s, Hawthorne, has patented these compounds for extraction purposes only in 1997 [10, 11]. The long-standing cooperation of NRI and RI, during which the cobalt bis(dicarbollide) technology was developed up to the stage of plant testing, is as a rule not mentioned in the subsequent Russian and American literature. During the years, several review articles and reports on the extraction with cobalt bis(dicarbollide)s appeared. Among early reviews, a paper of Makrlik et al. [12] should be mentioned. Some properties of cobalt bis(dicarbollide) technological systems during the cold war period appeared in two American reports [9, 13]. Cobalt bis(dicarbollide) technology as developed in Russia until 1992 was summarized in the Russian monograph [14]. The more recent state (1994) was also briefly described by Kyrs [15]. The huge body of material concerning the development of cobalt bis(dicarbollide) extractants due to NRI and RI cooperation was collected and transferred for public attention to DOE in the form of a report of NRI in 1993 [16], but again this report apparently disappeared in the black hole of early post–cold war era. Important contributions to the cobalt bis(dicarbollide) chemistry were made after 1990 by the Japanese Atomic Energy Research Institute (JAERI) and LANL as referenced later in this review. A significant milestone in the development of technologies with cobalt bis(dicarbollide)s was the start of the plant for the fractionation of highly radioactive waste, for the separation of 137Cs and 90Sr from defense waste, in Russia in August 1996 [17, 18]. This is so far the only instance of a current extraction plant scale aimed at retreatment of the radioactive waste running in the world (with the possible exception of the extraction of the above two elements with crown ethers also accomplished in Russia). Intensive research on still more elaborated technological improvements of the process with chloro-protected cobalt bis(dicarbollide), including proposals of new synergists, new solvents, and improved stripping mixtures is documented in many papers originating after the time of the first industrial application from the RI. New metal bis(dicarbollide)s and other boron derivatives are studied in the frame of current EC projects; these will be reported in the present review. The new methods of synthesis, largely employed at IIC and other European universities, are inspired by the success of the cobalt bis(dicarbollide) technology. They generally aim at determination and synthesis of new derivatives with incorporated selective function groups. This approach can ultimately lead to finding new properties of classical extractants, like P苷O based compounds or crown ethers. In fact, binding the selective group to the cobalt bis(dicarbollide) moiety leads to negatively charged ions, in contrast to the neutral reagents, and completely different behavior can be expected. Finally, new ecologically suitable nonpolar solvents, based on combinations with n-dodecane or isopropylbenzene, are now studied at NRI, Rez. The convenient

Copyright © 2004 by Marcel Dekker, Inc.

246

Rais and Grüner

mixtures may, contrary to previous expectations, dissolve chloro-protected cobalt bis(dicarbollide) and lead to solvent mixtures of a third generation. Various newly proposed selective extractants for specific purposes in synergetic mixtures with H+12- (see Table 1 for its identity) are tested at NRI. Returning to the most recent reports, cobalt bis(dicarbollide) technology has been proposed in the United States as a part of future planned nuclear fuel retreatment there. The isolation of two major heat sources, 137Cs and 90Sr, from the waste can substantially improve also the economy of the planned nuclear waste repository, since cooling requirements of the huge volumes of repositories should be drastically reduced [19]. Interestingly, in the report [19] no reference is made either to the initial NRI-Russian nor to contemporary DOE-Russian studies. In spite of the omission, this proposal testifies that cobalt bis(dicarbollide) technology is now rooted firmly into the worldwide technical state of art. A detailed review on extractions in systems with ion pairs was recently published by Moyer and Sun [20]. This review is in certain aspects complementary to the present one, as the subject of extraction by hydrophobic anions and by cobalt bis(dicarbollide)s was to some extent treated there. The main body of the review [20] was, however, devoted to the development and testing of the authors’ model devised and aimed at theoretical calculation of the selectivity of the extraction of alkali metal cations, a goal different from the present one.

B. Scope and Aims of the Review In organizing this survey, we tried to include the main new experimental and theoretical knowledge that may be of use for a potential reader. New, nonpublished results are also included here. We do not insist on the completeness of the review, providing space for a reader to search by himself some concrete information of interest that is not considered here as of primary importance. Several areas had to be omitted in the review, like the X-ray and structural data for new boron compounds, or detailed treatment of extractions with other than cobalt bis(dicarbollide) anions. These broad subjects surpass the scope of the review and could be a matter for other review articles. There exist many analytical procedures based on the cobalt bis(dicarbollide) extractants; these are treated here only in a general review style. The studies of radiation stabilities of cobalt bis(dicarbollide)s and solvents used at 1970 and 1980, though they might be important form a theoretical point of view, were practically superseded by newly proposed solvents. Thus, this area is not covered in detail. The aims and scope of this review are as follows: 1. 2.

Detailed description of synthesis of various metal bis(dicarbollide)s and other boron extractants together with their general and extraction properties Chloro-protected cobalt bis(dicarbollide) technology for the treatment of radioactive waste solutions

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions 3.

247

Description of other published technological and analytical uses of cobalt bis(dicarbollide)s

The present review is complemented by another review appearing in this volume (Chapter 6) that is devoted to the principles of the extraction of electrolytes [21]. The mechanism of extraction and physicochemical principles described there apply fully to the extractions with metal bis(dicarbollide) anions.

II. SYNTHESIS AND PROPERTIES OF METAL BIS(DICARBOLLIDE)S AND OTHER CLUSTER BORON COMPOUNDS AIMED FOR EXTRACTION PURPOSES A.

General Properties of Metal Bis(dicarbollide)s and Other Cluster Borate Anions

This section is focused, almost exclusively, on a small segment of boron cluster species based on the stable 12-vertex anionic metallaborate, borate, and carbaborate compounds with icosahedral closo structures which have been designed, synthesized, and tested with the aim of their subsequent use in liquid-liquid extraction. This category of anion-forming compounds ranges now from the parent metal bis(dicarbollide)s through their boron and carbon substituted derivatives to mixed sandwich compounds and simpler closo anions such as the substituted [CB11H12]- and [B12H11NH3]- derivatives. For their schematic structures see Fig. 1 and Tables 1–11.

Figure 1 Schematic structures of the univalent anions most studied for extraction purposes, including the cage numbering schemes. (a) Parent cobalt bis(1,2-dicarbollide)(1-), (b) isomeric parent cobalt bis(1,7-dicarbollide)(2-), (c) 1-amine-closo-undecahydro dodecaborate- ion, (d) closo-1-carba-dodecahydro dodecaborate- ion.

Copyright © 2004 by Marcel Dekker, Inc.

248

Rais and Grüner

Table 1 B-halogen and B-Alkyl Substituted Cobalt Bis(1,2-dicarbollides)

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions Table 1 Continued

Copyright © 2004 by Marcel Dekker, Inc.

249

250

Rais and Grüner

Table 1 Continued

# DCs measured at NRI under following conditions: 0.01 M solution in nitrobenzene, 1 M HNO3; approximately the same value of distribution coefficient is obtained for all halogenated derivatives, regardless to the number of substituents. NA=data not available. References containing also extraction results are denoted with *. When no references for extraction data is given, then the results are from Ref. 256.

The general chemistry of these anions has already been the subject of extensive reviews and monographs. Boron-cluster compounds can be defined as threedimensional aggregates of {BH} cluster units interconnected by 2-electron 3-center B–B–B bonds and classical 2-electron 2-center B–B bonds, the former being the result of the electron-deficient nature of boron in borane molecules [22, 23]. The corresponding borane molecules are composed of a defined number of {BH} vertices (n) that are arranged as triangular facets of a deltahedral cage. The geometry of the cluster is dependent on the number of boron vertices and cage electrons [23]. Each of the {BH} vertex units provides 3 orbitals and only 2 electrons to the cluster bonding scheme. This results in electron deficiency, formation of 3-center bonds, and an extensive electron delocalization over the whole cluster area. The most stable 12-vertex series, which will be discussed here, has 26 cage electrons present in the cage orbitals and adopts an icosahedral geometry. This class belongs the wide series of cluster borate [BnHn]c- anions, in which the values n=12 and c=2 fit for the particular icosahedral arrangement. Notional replacement of {BH}- cluster units by the isolobal {CH} group (providing 3 skeletal electrons) then leads to carboranes of the basic structural formulas [CmBnHn]c m (m is 1 or 2 for known carbaboranes of the closo series). Substitution of the {BH}- moiety by other main groups atoms such as S, P, or N leads to heteroboranes or mixed carbaheteroboranes, etc. An insertion of a metal atom or metal moiety (generally regarded as more electron deficient than boron) into a carborane framework generates

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

251

metallacarboranes. The basic geometrical and electron counting considerations for closo cage metallaboranes and metallacarbaboranes reflects those used for carboranes and heteroboranes. Much of the early and present progress in the metallacarborane area emerged from the recognition by M.F.Hawthorne in 1965 [1, 24–26] that the frontier orbitals of the [C2B9H11]2- nido dicarbaborane dianion (“dicarbollide” anion) open face are similar to those of the cyclopentadiene ion [C5R5]-. This similarity has been subsequently experimentally verified by synthesis of a broad range of mixed sandwich and full sandwich dicarbametallaboranes. These have been followed by other classes of cluster boron compounds, such as monocarba-, tricarba-, tetracarba-, and recently also pentacarbametallaboranes and heterocarbametallaboranes possessing closo-, or open cage nido-, and arachno-structures (not discussed here), including structural types lying on the borderline. The carbametallaboranes were described in hundreds of original studies that have been summarized in extensive reviews and monographs [27–32]. This review will cover only a small section of this extensive area. In principle, bis-icosahedral metalla-complexes containing two dicarbollide ligands [(C2B9H11)2M]c- [c=2 for M(II), c=1 for M(III) and c=0 for M(IV), respectively] are analogues of metallocene complexes that are well known from organic chemistry and are widely used as catalysts in polymer synthesis, supramolecular chemistry, medicinal use, etc. With respect to central metal atom complexation, the dicarbollide [C2B9H12]2- anion serves as a 6-electron donor η5 bonding ligand, preferring low spin complexes. Due to their formal divalent negative charge, the dicarbollide ligands are known to stabilize complexes with high metal oxidation states. The principal differences between the chemistry of metallocene and dicarbollide-metal-sandwich species lie in the very stable space filling “peanutlike” structures formed by the later, which can be described in terms of two fused icosahedra units sharing a common metal vertex. The metal bis(dicarbollide)s possess, in the majority of known cases, a negative charge delocalized over the large surface of the molecule. This and the particular character of bonding in the cluster lead to significantly enhanced thermal, chemical, and electrochemical stabilities.

B. Cobalt Bis(dicarbollide)s 1. Nomenclature A unique place in the design of the borate anion extraction agents has been occupied from the early beginning by the [closo-commo-(1,2-C2B9H11)2-3-Co(III)]- anion (1) and its derivatives (see Fig. 1). It is appropriate to make here some remarks on the nomenclature and cage numbering of this and relates species, due to discrepancies often seen in the literature. The descriptor numbering scheme is shown in Fig. 1; the cage is numbered starting from the carbon positions. If the deltahedral ligands are not substituted, the rotation barrier imposed only by partial the δ+ charge on the carbon atoms is low, and both ligands can almost freely rotate

Copyright © 2004 by Marcel Dekker, Inc.

252

Rais and Grüner

in solution around the axis connecting the atoms B10-Co-B10. In this case, clockwise spiral deployment has been preferred, according to IUPAC nomenclature. However, special cases may apply: Bulky groups at the rim of the pentagonal ligand planes can sterically hinder free rotation, or an exo-skeletal bridge could stop it definitely. In such cases, mesomeric and racemic forms can arise as has been pointed out many times in the literature; for example, see Refs. 33–37, where deeper discussion appears. If the racemic form is resolved into enantiomers, absolute configuration of the σ-enantiomer corresponds to a dextrorotary spiral whereas the ρ-enantiomer applies to the laevorotary alternative. In this case, both orientations should be considered equal, as has been discussed several times in papers from this group [36, 37]. The ACS [38] and IUPAC [39] nomenclature recommendations for the anion 1:3,3'-commo-cobalta-bis(undecahydro-1,2-dicarba-closo-dodecaborane)-(1-)ate seem too complex and cumbersome for practical purposes. For this reason, semitrivial or trivial names have been proposed and can bee found more frequently in the literature. The semitrivial names bis(1,2-dicarbollido)-cobalt(III) (1-)ate, cobalta(III) bis(1,2-dicarbollide), or cobalt(III) bis(1,2-dicarbollide) were proposed by Hawthorne and associates [26]. Trivial or technical names have often appeared in the research reports, patents, and even in the open literature, such as COSAN (abbreviation from Cobaltacarborane Sandwich Anion) proposed by the Institute of Inorganic Chemistry for technological use, cobalt dicarbollide, CDC or simply dicarbollide. The latter name, however, can be confused with the 11-vertex dicarbaborane anion without metal, and thus is not recommended. The term cobalt bis(dicarbollide) seems the most widely used in the recent literature and will be used also within this review. From geometrical considerations it follows that 45 positional isomers are possible when considering the presence of four carbon atoms in the two dicarbollide ligands of the [(C2B9H11)2-3-Co]- ion. Surprisingly, one sole isomer has been prepared and characterized in this anionic cobalt bis(dicarbollide) series [26]. This is the 2,2'-commo-cobalta-bis(1,7-dicarba-closododecaborane)-(1-)ate ion, or cobalt(III) bis(1,7-dicarbollide) (2). This isomeric alternative is characterized by nonadjacent (meta) positions of the carbon atoms, lying still in the proximity of the metal atom. For its schematic structure and cage numbering scheme, see Fig. 1b. This limited number of anionic isomers is in deep contrast to the mixed-sandwich series of the neutral [3-(␩5-C5H5)CoC2B9H11] complexes, where a large variety of the positional isomers has been reported [40]. 2.

The Parent Cobalt Bis(dicarbollide) Anions: Synthesis and Properties

The chemistry of the [(1,2-C2B9H11)2-3-Co]- (1) and [(l,7-C2B9H11)2-3-Co]- (2) anions [24–26] has been continuously developed over nearly four decades. It has been already reviewed several times [27, 29] most recently and comprehensively in 1999 [41]. The chemistry of the former species is certainly the most studied among the carba metalla carboranes. The fact that the ion 1 is diamagnetic is an

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

253

advantage, which allows for convenient characterization of its derivatives by 11B, 13 C, and 1H NMR techniques. The main properties, including the m.p., 11B, 1H, UV, and IR spectroscopic and electrochemical data are summarized in Ref. 42. Cobalt bis(dicarbollide) belongs to a class of low nucleophilic and low coordinating anions [43, 44]. The X-ray determined molecular structure of Cs1 was published soon after its first synthesis [45]; however, the carbon positions were fully refined later in the structure of Et3NH1 [46]. The molecular size of this ion is relatively large, the mean distances B(10)-B(10') and B(4)-B(7) and C(2)-B(8) cluster positions being 7.820(8), 2.871(6), and 2.787(5) Å, respectively (average values based on structural data of 31 nonbridged compounds from the Cambrige Crystallographic Data Centre). The terminal hydrogen atoms of 1 have a hydridic character. Along with charge delocalization over the surface, this is apparently the main cause of the high degree of dissociation of the very strong free conjugate acids, and the unique hydrophobicity of all cobalt bis(dicarbollide) derivatives. A characteristic feature of the bis-icosahedral cobalt bis(dicarbollide) is good solubility of its free conjugate acids and most of their salts in medium polarity solvents like ethers, nitro-solvents, halogenated solvents, etc, to which they can be extracted from an aqueous phase. The salts with bulky cations are sparingly soluble in water [42, 47]. These properties were recognized by Plesek already 30 years ago [48]. The unique importance of 1 and 2 in the design of extraction agents depends on the extraordinary chemical and thermal stability of 1 and 2 due to their “pseudoaromaticity” and to the completely filled electronic shell of the Co(III) cation which is furthermore sterically shielded by two bulky dicarbollide ligands. The central cobalt atom is chemically inert and stable towards a nucleophilic attack. Furthermore, the oxidation-reduction potentials are high [42, 49]. Reduction of Cs1 with 1 equivalent of Na(Hg) amalgam or Cs metal proceeds on the central atom without cage decomposition; only the cobalt(II) dianion [Co(II)(1,2C2B9H11)2]2- was isolated, which reverts back to 1 upon exposure to 0.5 equivalent of elemental iodine [27]. This was proven by the molecular structure of the solvated salt of this anion Cs2(DME)4[Co(1,2-C2B9H11)]2- determined by X-ray diffraction. In this particular case, according to some observations, agents such as RLi can also reduce the central cobalt atom [50]. In such circumstances, the lithium atom can penetrate between the two dicarbollide ligands close to the central atom, due to its small diameter. On the other hand, other authors have reported that lithiation of carbon vertices only proceeds when anion 1 is treated with BuLi [51]. The parent cobalt bis(1,2dicarbollide) anion 1 in the form of its Cs salt is stable toward HNO3 up to 2 M for several days contact. Beyond this concentration, the hydridic H atoms in positions B(8) and B(8') are slowly oxidized to -OH groups or substituted by the strongest nucleophile present in the system. Cobalt bis(1,2dicarbollide)s are exceptionally stable toward radiation. Experiments have shown that they survived radiation of 106 Gray per 24 h [52]. The cage is slowly degraded upon treatment with concentrated solutions of NaOH or KOH in protic solvents. The first detectable step is the degradation of

Copyright © 2004 by Marcel Dekker, Inc.

254

Rais and Grüner

one of the symmetry equivalent boron atom adjacent to carbon sites, B(6,6'). A cobalt dicarbaborane is formed which contains one open 11-vertex nido cage and one 12vertex closo cage [(1,2-C2B9H12)(7',8'-C2B9H11–3,9'-Co]2- sharing the common cobalt atom. The open-cage part of this ion can be further protonated to the [(1,2C2B9H12)(7',8'-C2B9H11–3,9'-Co]2- univalent anion, oxidized using H2O2 to the closo species [(1,2-C2B9H10)(7',8'-C2B9H11–3,9'-Co]-, or selectively degraded by FeCl3 to the a product missing another boron vertex [53]. Complete degradation to boric acid occurs upon a long-term contact. All reported synthetic strategies to the anions 1 and 2 are based on closo-1,2C2B10H12 (and closo-1,7-C2B10H12), (ortho- and meta-carboranes) [54, 55] as the starting materials; see Scheme 1. The icosahedral o-carborane cage undergoes one boron degradation in position B(3) (or symmetrically equivalent B(6) in the parent unsubstituted compound) when treated with a large variety of basic reagents with the formation of the [7,8-C2B9H11]- (dicarbollide) anion [56, 57]. Typically, NaOH or KOH in methanol or ethanol are used [58] for the unsubstituted compound, although numerous other reagents and conditions were designed using organic bases or the F- ion [59–69]. These are, in turn, particularly useful in the synthesis

Scheme 1 General synthetic route leading to cobalt bis(1,2-dicarbollide)- ion or neutral mixed 3-(␩5-C5H5)-3-Co-(1,2-C2B9H11) complexes. Parent as well as substituted compounds were prepared using this approach, under a large variety of particular reaction conditions.

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

255

of substituted and less stable ions. The [7,8-C2B9H12]- anion still contains a hydrogen bridge, which can be is in situ removed using either NaH (which correspond to the original procedure described by Hawthorne [26, 56, 57] similar to the procedure using BuLi [70]), or alternatively by KtBu [49, 71] in ether solvents. The ether solution of the divalent anion [7,8-C2B9H11]2- is then reacted with anhydrous CoCl2 at higher temperatures. The divalent anion [7,8-C2B9H11]2-, serving as the important intermediate in the synthesis, remained for a long time surprisingly poorly characterized until a recent detailed NMR and theoretical study [72] of its Li+, Na+, and K+ salts. In the synthesis of the parent cobalt bis(dicarbollide) and some other unsubstituted metalla bis(dicarbollide)s, aqueous conditions can also be applied using CoCl2 · 6H2O and 10 M KOH [25, 26]. In this case, an equilibrium between dicarbollide monoanion and dianion forms in the solution, which is sufficient for the formation of the metal complex. Typical yields of the cesium salt of the parent anion 1 are usually within 70–80%, regardless if anhydrous or aqueous procedures are applied. On the other hand, a simultaneous in situ degradation of the cage of 1 proceeds under conditions of the aqueous route, which in turn leads to a small yield of the “double-decker” sandwich with a divalent anionic charge [(C2B9H11)Co(C2B8H10)Co(C2B9H11)]2- (3) (see Table 1) as the main side product from the synthesis. Formation of this ion and a larger trimetallic “triple-decker” sandwich anion [(C2B9H11)Co(C2B8H10) Co(C2B8H10) Co(C2B9H11)]3- in strongly basic solutions was revealed and studied already in the early period of the metallaborane chemistry [73, 74]. The molecular structures of these anions were determined by X-ray diffraction studies [75, 76]. Synthetic routes to even higher congeners were investigated more recently [77]. Extraction properties of 3 have also been tested and found similar to that of parent cobalt bis(dicarbollide) 1 [78]. The isomeric cobalt(III) bis(1,7-dicarbollide) (2) (see Fig. 1) arises from base degradation of meta-carborane 1,7-C2B10H12 to the [7,9-C2B9H12]2- ion and metal insertion reaction [26]. Only a limited number of studies describing extraction properties of this otherwise even more stable species appeared in the literature. This has been undoubtedly influenced by the higher price of the m-carborane and its much slower and difficult degradation to the respective [7,9-C2B9H12]2- anion, which is the limiting reaction step in the synthesis. 3.

Substituted Cobalt Bis(dicarbollide) Ions: General Synthetic Methods

From the point of view of synthetic chemistry, the main characteristic feature of 1 (and 2) is the ability of these anions to undergo an easy substitution of its B(8,8') (or B(6,6')) terminal hydrogen atoms for a variety of nucleophiles L. This leads to the formation of the respective B(8)-L and B(8,8')-L2 (or B(6)-L and B(6,6')-L2) derivatives (or bridged structures) [41]. This is apparently the reason why literature reports on boron substituted cobalt bis(dicarbollide) compounds prevails over carbon derivatives of 1. This is in a contrast to the chemistry of the neutral closo-carboranes,

Copyright © 2004 by Marcel Dekker, Inc.

256

Rais and Grüner

where the situation is reversed [54, 55]. Many examples are reported in the abovementioned review [41] and in citations therein. The reaction proceeds either via an electrophilic substitution reaction path known from organic chemistry, or by a mechanism that has been defined as electrophile induced nucleophilic substitutions (EINS) [33, 79]. The later pathway is practically unknown for organic compounds, but can be considered as the typical reaction scheme within the boron cluster series and, in particular, for the anionic species. The basic principles of such reactions are as follows: an electrophile, eventually generated in situ, abstracts the hydridic terminal hydrogen atom from the most negative B–H vertex of the boron cluster position (B(8)-H and the symmetry-equivalent B(8')-H in 1) and the transient vacant B orbital then becomes attacked by the most nucleophilic or most electron-rich moiety from the environment. An important alternative way to substituted derivatives proceeds via degradation of already substituted carboranes and subsequent metal incorporation (see Scheme 1). It should be noted that this reaction path is rarely used for synthesis of boronsubstituted compounds, an is used only if special conditions apply, e.g., for radiolabeled [80–82] compounds or isomers not accessible by direct substitution reactions. The reason for less frequent use of this way is obvious: Anhydrous conditions are usually required to accomplish the metal incorporation and the reactions give typically lower yields of the resulting substituted dicarbollide anions; for example, see Ref. 80. The opposite is true for the synthesis of carbon substituted derivatives, where this reaction scheme has been considered until now as the main and the most reliable preparative route. Recently, a direct synthesis proceeding via lithiation of C–H bonds of the cobalt bis(dicarbollide) and subsequent reactions with alkyl halides was reported [51]. On the other hand, other authors failed to employ these reactions, at least for some more sophisticated substitutions [83]. 4. Boron Substituted Anions As has been discussed above, the B(8,8') skeletal positions opposite to the carbon atoms, and in the vicinity of the central metal atom, identified as the sites of the highest electron density [84], are prone to the attack by a nucleophile in the presence of the strongest Brõnsted or Lewis acid reagents that are able to abstracts B(8,8') hydrogen atoms. A reverse side of the easy reactivity of the B(8,8') positions is reflected in the eventuality of introduction of a hydroxy or other polar group upon longterm contact of the parent anion 1 with strong oxidizing aqueous acids. Presence of such polar groups increases the solubility of the resulting products in the aqueous phase and decreases the extraction efficiency. This was recognized very soon in the extraction process development, and hence the most reactive vertices were blocked by appropriate substitution. A large number of derivatives of the anion 1 with substituents at the most reactive B(8,8') positions have been prepared and studied for liquid-liquid extraction. The presence of substituent(s) has increased the chemical stability of the molecule substantially, especially in respect to attack by nitric acid and oxidation in the process. The simplest approach applied from the

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

257

beginning involved halogenation of the boron sites B(8,8') and also the vicinal positions B(9,9',12,12'). Later, aryl [85] and alkyl [86] substituents were introduced in these sites. More sophisticated substitutions within the series of compounds yield anions that simultaneously play the role of a selective metal scavenger that is able to tightly complex the target ions. Halogenated Cobalt Bis(dicarbollide)s. Most of the cluster borate and carborate anions are known to undergo halogenation reactions readily when treated by elemental halogens or with a variety of halogenating agents; for examples, see Refs. 43, 44, 87. These reactions were successfully applied for the protection of the most reactive sites of the cage of 1. As was found almost three decades ago, the oxidation of the cage 1 can be effectively prevented by chlorination leading to [(8,8',9,9',12,12'-Cl6-cobalt bis(dicarbollide) and the best extractants developed for the dicarbollide technology before 1995 belonged to the class of halogenprotected cobalt bis(dicarbollide)s. The halogenated derivatives (for schematic structures and basic extraction data, see Table 1) are very stable species, especially in acidic conditions. Most of them survive long-term contact with up to 10 M HNO3 without a noticeable decomposition [48, 42]. The substitution of the two most reactive positions by halogens is sufficient to protect the boron skeleton. Disubstituted derivatives were proven to be stable in 3 M HNO3 for over a 1-month period [88]. Nevertheless, the hexachloro protected anion 1 is the most widely used species due to its better accessibility and higher hydrophobicity compared to that of the parent compound and the less halogenated species. It should be pointed out that the character of B–X (X=Cl, Br, I) bonds in the cluster boron compounds is unique and differs dramatically from that known in organic chemistry or from simple (BX3) compounds and their organic derivatives; for some illustrative data, see Ref. 88. Although the B–X bonds are longer than that in BX3 their stability is considerably higher. Although quantitative data for dissociation energies are not available, experimental evidence proved this on a qualitative level. A good example of this experimental evidence is the low reactivity of the boron-halogen bonds in the halogen derivatives of the basic icosahedral anion. Halogen atoms from perhalogenated derivatives or cannot be abstracted from the cage even by treatment with alkali metals in liquid ammonia or THF [87]. This is even more valid for less halogenated derivatives. Replacement of the halogen by another nucleophile proceeds only exceptionally and under forced conditions. Iodine substituents of the periodinated or bromine atoms from perbrominated compounds can be partially replaced by the CN moiety under intensive UV irradiation, although less halogenated compounds do not react [87]. Similar observations have been made for the chloro and bromo derivatives of the cobalt bis(dicarbollide). Abstraction of the halogen atom was not possible even upon treatment by NaNH2 or RLi reagents [88]. The stability of the B–X bond is considered rather to be a consequence of kinetic and steric factors [88].

Copyright © 2004 by Marcel Dekker, Inc.

258

Rais and Grüner

A noticeable drawback arising from the electronic effect induced in the cage by the halogen substituent(s) is that the positions B(6,6') which lie in the vicinity of the carbon atoms become even more sensitive to degradation with strong bases than those of the parent anion 1 (see text above). In the first stage of such degradation reactions, species [(C2B8H7X3)(1,2-C2B9H8X3)-Co]2- are produced that are not sufficiently soluble in the organic phase and thus are not suitable for extraction purposes. From this point of view follows the limited suitability of halogen derivatives for treatment of strongly basic nuclear waste solutions. The first halogenation reaction of the anion 1 to its hexabromoderivative was reported by Hawthorne [26]. Chlorination, bromination, and iodinations were later studied in the 1970s and 1980s in more detail in connection with extraction process development carried out at the IIC and NRI groups [4, 7, 89, 90]. Studies of the physicochemical properties were done in cooperation with the Comenius University at Bratislava [52, 91–97] and later also by other authors [88]. In contrast, the first successful fluorination reactions were reported as late as 2000 [98]. The use of the mild fluorinating agent, 1-chloromethyl-4-fluoro-1,4-diazoniabicyclo[2.2.2]octane bis(tetrafluoroborate) (F-TEDA) in anhydrous acetone accomplished fluorine substitution of anion 1, yielding the B(8,8') difluoro derivative [(8,8'-F2-(1,2-C2B9H10)23-Co]- (4). The structure of the salt of the B(8,8') derivative 4 was determined by single crystal X-ray crystallography. The chlorination reactions of 1 with elemental chlorine in ethanol-tetrachloromethane, nitrobenzene-tetrachloromethane or acetic acid lead to stepwise chlorination at the boron atoms known to possess the highest electron densities, i.e., in the order B(8), B(8,8') B(8,8',9), B(8,8',9,9'), B(8,8',9,9',12), B(8,8',9,9’12,12') [91]. In spite of the fact that the reaction rate decreases with the number of introduced halogen atoms, the presence of the species containing up to nine chlorine atoms was observed in some technological samples recently studied by Electrospray M.S. methods [99]. According to the explanation given in the older literature [91], these highly chlorinated species result from poor reaction rate control in the early reaction steps. This can lead to a statistical substitution proceeding simultaneously at less favored skeletal sites, followed by the reactions proceeding at the usual positions. The stereochemistry of these derivatives remains unknown and it may be only assumed that positions B(10, 10') would be the probable substitution sites. The monochloro derivative [(8-Cl-1,2-C2B9H10)(1',2'-C2B9H11)-3,3'-Co]- (5) was prepared by direct chlorination in a CCl4-ethanol mixture, or by careful halogenation using ␥ irradiation of 1 in chloroform-benzene mixtures [91]. This is a distinguishable step, due to the decrease of the reaction rate upon introduction of the first chlorine atom [91]. Since the reactivity of the positions B(8,8') is substantially higher than that of the B(9,9',12,12') sites, the disubstituted anion [(8,8'Cl 2-(1,2-C 2B 9 H 10 ) 2 -3-Co] - (6) could be prepared in pure form by direct halogenation carried out in the same solvent and controlling carefully the volume of the introduced elemental chlorine [91]. Other authors reported that derivative 6 can be obtained as the single product, from direct halogenation carried out in THF

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

259

in presence of iron powder, or by heterogenous reaction of the tetramethylamonium salt Me4N1 with a NaOCl/HCl mixture in water (whereas the same reaction with other cations led to cage decomposition) [88], or by halogenation in THF using NaOCl/HCl as the halogenation agent. More reliable seems an alternative way, when the derivative 6 was obtained by halogenation of 1 with N-chlorosuccinimide [88]. The molecular structure of the Me4N6 was determined by X-ray crystallography [88]. The isomeric B(9,9') dichloroderivative [(9,9'-Cl2-(1,2-C2B9H8)2-3-Co]- (7) was prepared starting from halogenation of ocarborane to the 9-C2B10H11Cl cage by degradation and metal insertion reactions [91]. On increasing the chlorine content beyond two atoms, mixtures of several species are always obtained from direct halogenation. Especially, if an odd number of halogen atoms is introduced, isomeric and diastereoisomeric mixtures of each derivative would necessarily be produced, further complicating the isolation of pure species. Despite of this, isolation of almost pure trichloro derivative [(8,8',9Cl3-(1,2-C2B9H9)(1',2'-C2B9H10)2-3-Co]- (8) was reported as being done by direct chlorination in nitrobenzene-CCl4 under cooling followed by chromatographic separation of the products on a Sephadex LH-20 column [91]. Similar purification provided the tetrachloro derivative (9) [(8,8',9,9'-Cl4-(1,2-C2B9H9)2-3-Co]-. The isomeric [(9,9',12,12'-Cl4-(1,2-C2B9H9)2-3-Co]- derivative (10) was prepared via degradation of the 9,12-Cl2-1,2-C2B10H10 substituted carborane and subsequent metal insertion reaction [91]. On the other hand, the leading position in the extraction technology is governed by the hexachloro-substituted product [(8,8',9,9',12,12'-Cl6-(1,2-C2B9H8)2-3-Co](11) [89–91]. This derivative could be prepared in an almost pure state as a smallscale laboratory chemical [91]. However, for larger-scale chemicals, used for extraction purposes, only an average composition of six halogen atoms could be obtained. These materials are typically mixtures of derivatives containing 5, 6, and 7 chlorine atoms, including their respective isomers. These would be referred to as chloro-protected cobalt bis(dicarbollide) (12) (a technical abbreviation frequently used in Russian Federation and the United States is ChCoDic). All the above chloro derivatives were also obtained from radiolysis of 1 in chloroform-benzene or CCl4nitrobenzene solvent mixtures, depending on the dose absorbed [91]. The direct chlorination in acetic acid based on the patent [90] and performed under cooling till introduction of approximately two equivalents of chlorine and then at ambient temperature is currently in the use for its production by the main supplier of this compound (Katchem Ltd. Prague). Addressing the above-mentioned possibility of halogenation to higher stages, special precautions should be followed during the production of 12 in order to ensure the average chlorine content to correspond to 6. These should consist of careful application of the analytical and separation methods such as elemental analysis, 11B and 1H NMR techniques [91], isotachophoresis [93, 97], HPLC [100, 101], or preferably LC-MS, during the chlorination stages of the production and careful analysis of the final product.

Copyright © 2004 by Marcel Dekker, Inc.

260

Rais and Grüner

Bromination reactions proceed more slowly and the monobrominated anion [8-Br-1,2-C2B9H10)(1',2'-C2B9H11)-3,3'-Co]- (13) can be obtained as well as the dibromoderivative [(8,8'-Br2-(1,2-C2B9H10)2-3-Co]- (14) by direct halogenation in methanol [91]. The latter compound 14 was more recently prepared using the reaction with N-bromosuccinimide in THF [88]. Both derivatives 13 and 14 resulted also upon radiolysis of nitrobenzene-tribromomethane or tribromomethanemethanol solutions of 1 [52]. The isomeric derivative [(9,9-Br2-(1,2-C2B9H10)2-3Co]- (15) was prepared from 9-Br-o-carborane, similarly as compound 7 in the respective chlorinated series [91]. Also the pure isomer [(9,9',12',12-Br4-(l,2C2B9H9)2-3-Co]- of the tetrabromo derivative 16 was prepared using an indirect approach from 9,12-Br2-o-carborane [102]. Another isomeric derivative [(4,7,4',7'Br4-(1,2-C2B9H8)2-3-Co]- (17) was obtained recently from the bromo substituted dicarbollide anion [9,11-Br2-7,8-C2B9H10]-, the molecular structure of which was determined by X-ray crystallography [103]. Room or higher temperature bromination reactions with elemental bromine result in the hexabromoderivative [(8,8', 9,9',12,12'-Br6-(1,2-C2B9H8)2-3-Co]- (18) [91]. The cesium salt of this derivative was characterized by X-ray diffraction analysis [94]. Lower reaction rates of bromination generally mean that the technological samples of bromo protected cobalt bis(dicarbollide) (19) are usually better defined than the respective chloro protected species 12. Direct iodination of 1 in methanol or ethanol gives either the iododerivative [(8-I-1,2-C2B9H10)(1',2'-C2B9H11)-3,3'-Co]- (20) or the diododerivative [(8,8'-I2-(1,2C2B9H10)2-3-Co]- (21) [91], depending on the initial ratio of iodine to 1. The molecular structures of the Cs+20- and salt [95] and of the disubstituted derivative Cs+21- [96] were determined by single crystal X-ray diffraction. Although the direct iodination does not proceed to higher stages even under heating, the hexaiododerivative [(8,8',9,9',12,12'-I6-(1,2-C2B9H8)2-3-Co]- (22) was successfully prepared by iodination catalyzed with AlCl3 [91], or by reaction with ICl, which seems to be advantageous over the older procedure due to simpler reaction conditions without need of the catalyst and a good yield (92%) [86]. In addition to the above halogen-substituted anions, zwitterionic compounds have also been reported, where both dicarbollide ligands were connected by iodonium and bromonium bridges in the positions B(8,8') [(8,8'-µ-X-(1,2-C2B9H10)23-Co] (X= I, Br) (23, 24) [104]. These arise from AlCl3 catalyzed intramolecular cyclization reaction of the B(8) iodo or bromo derivative. The compounds 23 and 24 deserve to be mentioned here, due to a facile cleavage of the B–X–B ring proceeding with a large variety of bases. This can serve for further syntheses of the B(8,8') protected compounds bearing a metal selective group. One such example is given in text below (see Section II.4.c, compound 40). Cobalt Bis(dicarbollide)s Substituted at Boron Atoms with Alkyl and Aryl Groups. Direct alkylation of boron vertices was reported only recently. Reaction of anion 1 with neat CH3I catalyzed with AlCl3 produced the [8,8'-(CH3)2-(1,2C2B9H10)2-3-Co]- anion (25) [105]. A similar procedure leading to the formation of

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

261

[8,8'-(i-C4H9)2-(1,2-C2B9H10)2-3-Co]- (26a) can be found in the patent literature [10] along with that for the dibutyl analogue derived from the isomeric anion 2 series [6,6'-(i-C4H9)2-(1,7-C2B9H10)2-3-Co]- (26b) [11]. The later two anions were designed for extraction purposes, but no data are available. An indirect method, i.e., (PPh 3) 2PdCl2/CuCl-catalyzed cross-coupling procedure of iododerivatives with alkyl Grignard reagents, has been reported. Synthesis of [8,8'-(n-C10H21)2-(1,2-C2B9H10)2-3-Co]- (27) using the reaction of the diiododerivative 21 with decylmagnesium bromide in THF led to 76% yield of this anion, according to a recent U.S. patent [10]. A palladium-catalyzed crosscoupling reaction of hexaiododerivative 22 with CH3MgBr produced the hexamethyl derivative [(8,8',9,9',12,12'-(CH3)6-(1,2-C2B9H8)2-3-Co]- (28) obtained in 60% yield [86]. The search for more effective compounds with enhanced hydrophobic properties for Cs+ and Sr2+ extraction promoted synthesis of various aryl-substituted derivatives, where the cage positions were bridged with a variety of aromatic organic groups. The first compound of this class [8,8'-µ-1,2-C6H4)-(1,2-C2B9H10)2-3-CO]- (29) was prepared in low yield, already in 1972, by thermally induced reaction of aryldiazonium salts with Me4N1 [106], but without any intention of its further use. The molecular structure as determined by X-ray diffraction was described in a subsequent paper [105]. A more reliable procedure leading to the bridged derivative 29 in better yield, however, consists of an AlCl3-catalyzed reaction of Cs1 with benzene used as neat solvent [107]. Recent substantial revisions of this reaction led to the synthesis of a large series of compounds bridged in positions B(8,8') with various arylene substituents [85, 108]. Compounds with phenylene 29, tolylene [8,8'-µ-(CH3-C6H3)-(1,2-C2B9H10)2-3-Co]- (30), ethylphenylene [8,8'-µ-(CH3CH2C6H4)-(1,2-C2B9H10)2-3-Co]- (31), o, m, p-xylylene [8,8'-µ-(CH3)2-C6H3)-(1,2C2H10)2-3-Co]- (32a, b, c), biphenylene [8,8'-µ-(C6H5-C6H4)-(1,2-C2B9H10)2-3-Co](33) and tetraline (34) bridging moieties were synthesized (see Scheme 2). For schematic structures and extraction data see Table 2. Reaction with naphtalene

Scheme 2 Synthetic procedure leading to arylene bridged and arylene bis-bridged cobalt bis(dicarbollide) anions. A mixture of both respective types always results, which should be subsequently separated.

Copyright © 2004 by Marcel Dekker, Inc.

262

Rais and Grüner

Table 2 Arylene-Substituted Cobalt Bis(1,2-dicarbollides)

0.01 M extractant solution in respective solvent indicated in exponent, 1 M HNO3, NPOE—nitrophenyl octyl ether, NB—nitrobenzene. #NRI—unpublished results. References containing extraction results are denoted with *.

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

263

carried out in cyclohexane as the solvent resulted in an interesting rearrangement of the aromatic ring providing compound, in which the dicarbollide ligands were interconnected via an unexpected 3-carbon atom bridge [8,8'-µ-CH2-C9H6)-(1,2C2B9H10)2-3-Co] (35) [109]. Also, the compound containing a nonbridging phenyl substituent at B(8) position [8-(C6H5-1,2-C2B9H10) (1,2-C2B9H11)-3-Co]- (36) was obtained in moderate yield from dimethyl sulfate—induced reaction of 1 with benzene [110]. The molecular structures of 35 and 36 were determined by X-ray diffraction [109, 110]. Revision and extensions of the AlCl3-catalyzed reactions along with careful product separations and characterization by modern NMR, mass spectrometric and crystallographic techniques emerged in the discovery of a novel promising class of 4,8',8,4'-R2-bis(arylene)- bridged cobalt bis(dicarbollide)s [4,8',4’8-µ(C6H4)2-(1,2-C2B9H9)2-3-Co]- (37), [4,8',4’8-µ-(CH3C6H3)2-(1,2-C2B9H9)2-3-Co] (38) and [4,8', 4’8-µ-(CH3CH2C6H3)2-(1,2-C2B9H9)2-3-Co]- (39) [85]. For the reaction path, schematic structures and extraction data see Scheme 2 and Table 3. Reaction conditions have been optimized to give up to ca. 54% yields of these interesting compounds [108]. The compounds with arylene substituents 37 to 39 were the subject of relatively extensive extraction tests. The basic member of the series 37 proved especially to have excellent complexation properties and extraction selectivity for the cesium cation, surpassing markedly the extraction ability of chloro protected cobalt bis(dicarbollide) (12). A significant advantage of this class of extraction agents lies in their reasonably high solubility in aromatic solvents (toluene, xylene, etc.), provided that some aromatic sulpho compounds, designed and used as solubilizers, are added to the organic phase (see Section VII.D) [111]. X-ray studies of the Cs+ complex of the anion 37 proved, that an angle of 72° between planes of phenylene substituents is favorable for a tight Cs+ complexation [108]. The distribution ratio of Cs+ has been found so high that it imposed a consequent problem in the stripping. As in the case of halogen-protected anions, this could be accomplished by nitric acid of high concentration. On the other hand, Table 3 Bis-Arylenelene Bridge Substituted Cobalt Bis(1,2-dicarbollides)

0.005 M extractant solution in nitrobenzene, 0.5 M HNO3. X, 0.005 M extractant in 0.4 M DEPSAM solution in toluene. References also containing extraction results are denoted with *. When no reference for extraction data is given, then the results are from Ref. 256.

Copyright © 2004 by Marcel Dekker, Inc.

264

Rais and Grüner

stripping with concentrated acid is not the optimal solution, since these compounds are more sensitive to the oxidation effect of nitric acid than chloro protected cobalt bis(dicarbollide) 12. Boron-Substituted Cobalt Bis(dicarbollide)s Functionalized with Metal Complexing Groups. A typical procedure for lanthanide and actinide Mn+ cation transfer from an aqueous solution into a low polarity organic phase uses tight complex formation with neutral ligands. In the known schemes employing organic ligands, metal is surrounded with electron donor atoms of the uncharged ligands like crown ethers, malonamides, phosphine oxides, CMPO (N,N-dialkyl carbamoyl methyl dialkyl phosphine oxides), or modified calixarenes, able to form strong, hydrophobic, but still n+ charged complex. In this case, hydrated nitrate anions have to be transferred across the interface together with the target cation. A new approach was recently conceived, aimed at the development of effective separation methods for Sr2+, lanthanides, and minor actinides, i.e., long-lived nuclides, from highly acidic solutions of high level radioactive waste (HLW). Several feasible synthetic ways were found that enabled the provision of anion 1 with cation ligating groups from the above series. Selective groups containing electron donor atoms (e.g., oxygen, nitrogen, and phosphorus) were successfully bonded to the 1 cage by covalent bonds. For schematic structures and extraction data of this class of new compounds, see Tables 4–7. Several new compounds of this type proved effective in transferring the target radionuclides from 1 M HNO3. The first attempt in this direction was made in the early 1990s when the [(8(C5H11-(CH2CH2O)2-1,2-C2B9H10)(8'-I-C2B9H10)-3,3'-CO]- anion (40), was prepared from the [(8-(CH2CH2OCH2CH2O)-C2B9H10)(8'-I-C2B9H10)-3,3'-Co] zwitterion (41) and tested for extraction properties [78]. The disubstituted intermediate compound 41 formed upon treatment of the iodonium bridged derivative 23 with dioxane [104]. The dioxane ring of 41 contained an oxonium oxygen and could be opened by NaOC5H11 to produce anion 40. The compound 40 was tested for strontium extraction, however without observation of extraction enhancement in respect to the corresponding synergetic mixtures [78]. A similar synthetic method was applied in the synthesis of a subsequent anionic series with covalently bonded metal ligating groups. This consisted of the ring opening of the 8-dioxane-1 [8-O(CH2CH2)2O)-C2B9H10)(1’2'-C2B9H11)-3,3'-Co] zwitterion 42. This interesting bipolar derivative became available on a larger scale only recently (based on results of the IIC group) [109] and proved to be a versatile synthon. As in the case of 41, the dioxane ring containing an oxonium oxygen atom can be easily cleaved by almost any nucleophile of choice [112–115] producing the species functionalized with various end groups attached to the hydrophobic anion 1 via a diethylene diglycol chain (see Scheme 3). Similar kind of reactions provides also the recently reported compound containing tetrahydropyrane ring [116].

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

265

Table 4 Cobalt Bis(1,2-dicarbollides) with Functional Groups Attached via Diethyleneglycol Chain

0.05 M 43, 44, 45, 47, 48 in toluene, 0.01 M 49, 50 in nitrobenzene, 0.01 M 51 in dichloroethane, 1 M HNO3. X, 0.1 M HNO3. References also containing extraction results are denoted with *.

Copyright © 2004 by Marcel Dekker, Inc.

266

Rais and Grüner

Scheme 3 The flexible synthetic route based on the cleave of the dioxane ring of the zwitterionic derivative 42. The reaction proceeds smoothly with a majority of possible nucleophic reagents, as has been already described in several papers (see Refs. 83, 112–115).

A novel series of compounds of the general formula closo-[(8-X-(CH2-CH2O)21,2-C2B9H10)(1',2'-C2B9H11)-3,3'-Co]- (X=1-O-2-CH3O-C6H4 (43), 1-O-2-C6H5CH2C6H4 (44), 1-O-4-tC8H17-C6H4 (45), 1-O-3-CF3-C6H4 (46), P(O)(OC4H9)2 (47) and P(O)(OC4H9)(OH) (48) (see Table 4 and Scheme 3 for the schematic structures and the reaction procedure), including several others [83] resulted from this method and were subsequently tested for extraction purposes. The flexible synthetic methodology implied a possibility to modify and study the influence of the character of the end group and the whole substituent, including the steric strain of the chain on the metal bonding properties. Results were described in the recent article [113] and following papers (see below). This study allowed deeper understanding of the effects of the particular selective groups covalently bonded to 1. The molecular structure of the sodium complex of the anion 43 was determined by single crystal X-ray diffraction. The sodium atom in this case is tightly coordinated to five oxygen atoms of the spacer chain and the guaiacolyl terminal group of 43, one water molecule, and from the opposite side of the ligand plane the short B(8')-H-Na contact [2.26(3)Å] was found to be within a bonding distance. The cation is enveloped by a hydrophobic outer sphere composed of hydrophobic CH2 and CH groups of the organic substituent and B–H groups of the cobalta bis(dicarbollide)(1–) cage [113]. The most interesting derivative found in the above series seem to be the anion 48 that contains the terminal monoester functionality The presence of this phosphoryl end group, acting apparently as the second acidic centre in the molecule, is sufficient to increase substantially the transfer of the M3+ cations into an organic phase even from highly acidic waste solutions. It was shown experimentally that this ester group is resistant toward further hydrolysis to the respective end substituent. Neither alkaline nor acidic conditions succeded in giving the expected product in good yield. On the other hand, partial hydrolysis in the extraction system may possibly explain the observed sharp increase of the DEu extraction coefficients after several days contact with 1 M HNO3.

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

267

Anionic crown ether compounds belonging to this family of compounds resulted upon cleavage of the dioxane ring of the zwitterion 42 with various sodium oxymethylcrowns differing in crown ether ring size [83, 114] (see Table 4). Three compounds [(8-(15-crown-5-CH2-O)-(OCH2CH2)2-1,2-C2B9H10)(1',2'-C2B9H11)3,3’Co]- (49), [(8-(18-crown-6-CH2O)-(OCH2CH2)2)-(1,2-C2B9H10)(1’2'-C2B9H11)3,3’Co] -, (50) and [(8–21-crown-7-CH2-O)-(OCH2CH2)2-1,2-C 2B9H10)(1’2'C2B9H11)-3,3’Co]- (51)] were prepared in high yields. These species represent the first examples where regular, oxygen atom-containing crown ether rings were covalently bonded to the cage of the hydrophobic anion 1. A second series of these derivatives, prepared for comparison (see Table 5), resulted upon deprotonation of the [(8,8'-(OH)2-(1,2-C2B9H12)2-3,-Co]- (52) anion by NaH in THF and subsequent reaction with p-toluenesulfonyl esters of the hydroxymethyl crown ethers [115]. Only compounds with single crown ether substituents could be isolated in preparative amounts from the respective reactions. Two compounds of this class [8-(15-crown-5-CH2O)-1,2-C2B9H10)(8'-HO-1',2'C2B9H10)-3,3'-Co]- (54) and [8-(21-crown-7-CH2O)-1,2-C2B9H10)(8'-HO-1',2'C2B9H10)-3,3'-Co]- (55) (Table 5) were prepared and adequately characterized. The molecular structure of the Cs+ complex of the species 54 was determined by single crystal X-ray diffraction. This structure exhibited coordination of two cesium atoms within two crown-cobalta bis(dicarbollide) ligands. Each Cs+ atom is coordinated to five oxygen atoms of one 15-crown-5 ring and to two O atoms of the second unreacted OH moiety in the position B(8') of the two ligands, thus forming a distorted square planar arrangement of two Cs+ cations and two O donor atoms. The hydrophobic core of the resulting complex, composed of the hydrophobic

Table 5 Cobalt Bis(1,2-dicarbollides) with Crown Ether Moieties Bonded via Shorter Chain

0.01 M extractants in nitrobenzene, 1 M HNO3. References also containing extraction results are denoted with *.

Copyright © 2004 by Marcel Dekker, Inc.

268

Rais and Grüner

anions and crown ether CH2 groups, is directed outwards of the complexed metal cation. The oxygen atom included in the -CH2-O-B(8) spacer does not participate on the metal complexation but its connection by intramolecular hydrogen bonds to the H–O–B moiety supports further the stability of this arrangement. The bonding of crown ether groups via a longer chain in compounds 49 to 51 has proven more favorable for Cs and Sr2+ extraction as was described in Ref. 114. The distribution coefficients for Cs+ and Sr2+ at higher acidities were found to be comparable to that observed for the synergetic mixtures of 19 and crown ethers, although a slight enhancement in Sr2+/Na+ selectivity was observed for covalently bonded species. A promising family of boron-substituted derivatives of 1 of the general formula [(8-CMPO-CH 2 -CH 2 O) 2 -1,2-C 2 B 9 H 10 )(1’2'-C 2 B 9 H 11 )-3,3'-Co] - [CMPO= Ph2P(O)CH2C(O)-NR, R=C4H9 (56), C12H25 (57), CH2C6H5 (58)] were prepared using a two-step procedure [115] (see Table 6). The first step consisted of the cleavage of the 8-dioxane-cobalt bis(dicarbollide) 42 ring by the respective primary amine. Subsequent reaction of the resulting amino derivatives with the highly reactive nitrophenyl ester of diphenyl phosphoryl acetic acid resulted in the

Table 6 Cobalt Bis(1,2-dicarbollides) with CMPO-like Groups Attached via Diethyleneglycol Chain

0.01 M extractant in toluene, 1 M HNO3. References also containing extraction results are denoted with *.

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

269

anticipated CMPO-substituted products in high yields. The molecular structure of the supramolecular charge-compensated complex Ln(58)3 · 3H2O was determined by single crystal X-ray diffraction. The crystallographic results illustrated the capability of this anionic ligand to completely displace anions from the primary coordination sphere of the lanthanide cation. The Ln3+ cation is tightly coordinated by six oxygen atoms of the CMPO terminal groups (two from each ligand) and by three water molecules completing the metal coordination number of 9. The atoms occupying the primary coordination sphere form a tri-capped trigonal prismatic arrangement (see Fig. 2). Exceptionally high liquid-liquid D Eu distribution coefficients were observed for all three members 56 to 58 of this series.

Figure 2 Schematic drawing based on the molecular structure of Ln3+ complex Ln(58)3. 3H2O (for details see Ref. 115). Three functionalized cobalt bis(dicarbollide) ligands 58 lie in the coordination sphere of the metal ion, being bound to the cation via bidentate CMPO substituents. Hydrophobic groups are directed outward from the metal binding region. This may account for the high extraction efficiency of this class of compounds.

Copyright © 2004 by Marcel Dekker, Inc.

270

Rais and Grüner

In connection with the above CMPO-like family of derivatives, another, substantially different derivative [(8,8'-(Ph2P(O)CH2(CO)N)100 for all three CMPO substituted cobalt bis(dicarbollide)s. The shape of the dependence of DEu on the aqueous acidity was in all cases the classical one for the synergetic mixtures of CMPO compounds with simple cobalt bis(dicarbollide)s, i.e. a monotonous decrease with acidity. Hence, the classical behavior of the pure CMPO with a maximum does not exist in this case. The extracted particle extracts with the derivatives in a form similar to the ion pair extracted from the synergetic mixture. From the extraction results it followed that practically two 56- anions bind to one Eu3+ cation, or in other words that sterical hindrance does not allow all three anions of 56- to enter the vicinity of the cation in the solution. This is in contrast to the solid phase measurements; see Section II.

Copyright © 2004 by Marcel Dekker, Inc.

306

Rais and Grüner

The ligand 56 is a very powerful extractant of Eu3+ and stripping into an aqueous phase may be difficult. Thus, from 13 possibilities of back-extraction only some have led to an acceptable result, namely a mixture of 0.1 M ammonium oxalate or a tertiary ammonium phosphate with 5% pentasodium salt of diethylene triamine pentaacetic acid, DTPA [135].

VI. CHLORO-PROTECTED BIS(DICARBOLLIDE) TECHNOLOGIES FOR EXTRACTION OF FISSION PRODUCTS AND ACTINIDE CATIONS FROM R ADIOACTIVE WASTES A.

Early Development of the Process

The chloro-protected cobalt bis(dicarbollide) process of extraction of 137Cs and 90Sr was developed to the stage of plant testing as a result of the mutual cooperation between the RI, St. Petersburg, and NRI, Rez, during the years 1975–1985. This early development demanded considerable effort from both sides since a quite new type of process was tested with a new and exotic reagent. From the Russian side, complete installation of hot cells in Gatchina and subsequently a commercial line at the enterprise Mayak were provided for the purpose. From the Czechoslovak side, a full study of the chemistry of the process was performed. Several Czechoslovak universities were engaged in the research: the Universities of Bratislava and Brno and the Technical University of Prague. The voluminous reports of the activities are cited in a preceding review [16] and contained in detail in two 5-year collaborative reports [146, 147]. The studies concerned the choice of suitable diluent for the chloro protected cobalt bis(dicarbollide), evaluation of the best PEG for the purpose, chemical and radiation stabilities of the reagents and solvents, fire and explosion hazard assessments, influence of possible contamination of the extractant by TBP on the performance, modeling of the process—including the cascade calculations, determination of extraction isotherms at various combinations of components, developing the criteria for checking the purity of the supplied H+12- and testing the individual batches of the product, methods of analysis of radionuclides (90Sr) and of the composition of the extractant, and several other questions. For the calculation of the equilibria, the respective constants determined from the distribution of microamounts of radioactive tracers of Na-, Rb+, Cs+, Sr2+, and Ba2+ were calculated by the program EXTRIT (See Chapter 6). From the values determined for 1, 4, and 3 M HNO3, only the last set is reported here. The solvent was 60 vol% of nitrobenzene +40% CCl4 with 1 vol% of Slovafol 909 added, 0.06 M H+12- was used, all at 25±2°C. The logarithms of the calculated constants are given in parentheses after each cation in the following order: Kex(Mz+, zB-)/Kex(MLz+, zB-)/β°1(MLz+)/(βa1(MLz+) (see notation in Chapter 6): H+(1.60/5.48/3.88/0.63), Na+(1.24/6.30/5.06/2.10), Rb+(3.78/7.00/3.22/1.60), Cs+(4.32/6.70/2.38/1.58),

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

307

Sr2+(3.70/11.85/8.15/~1.71), Ba2+(4.91/13.04/8.13/~1.97) [222]. Contrary to data in Table 12, no double complexes of the type ML2 were detected, but complexation in the aqueous phase was taken into account. The above values were used for the calculation of extraction isotherms for modeling the extraction cascade. The calculated data agreed reasonably with experiment for such cases as, e.g., extraction of micro-amounts of Cs+ in the presence of macroamounts of Ba2+, etc. The experimental tables of recorded data were used by Sraier [223] for the modeling of the cascade. These cooperative efforts provided a firm basis for further testing. Present technological processes of the kind are still based on the previous findings of the Czech scientists concerning the nearly “ideal” properties of cobalt bis(dicarbollide) regarding its properties of high hydrophobicity and superacidity [2]. Added to this, the discovery of the complexing properties of PEGs toward Sr2+ made it possible to envisage a viable technological process based on this reagent. Two findings related to the chloro-protected cobalt bis(dicarbollide) are of general interest. One is that it is extremely stable on long-term contact with even concentrated nitric acid. The other is that during radiolysis of the 1 anion in organic solvents containing as a component a bromo- or chloro-substituted alkane radiational synthesis of a halogenated cobalt bis(dicarbollide) anions takes place instead of decomposition of the anion, e.g, Ref. 224. In these aspects of stability, the anion diametrically differs from usual carbon based organic reagents and this property was the main reason of its broad application. In the development of the process, particular attention was paid to the choice and testing of a proper stripping agent for both Cs+ and Sr2+. The original choices included n-propanol in nitric acid, diisopropylether, urea, and ammono-complexes of zinc, ammonium salts, and other reagents, all tested at NRI. The final choice of using concentrated (~10–12 M) nitric acid was decided during the cooperation. Hydrazine nitrate in nitric acid for stripping Sr2+ and Ba2+ was proposed at NRI [146]. Such stripping agents permitted to devise a process in which no solids will remain after evaporation and chemical decomposition of the hydrazine into gaseous products. Because of lack of information in contemporary literature on the technological testing at those early stages (see Section I.A), abbreviated information is provided here. In the two following tests to be reported, 0.06 M H+12- solution in 60 vol% of nitrobenzene+40% of CCl4 was used as an extractant (1 vol% Slovafol 909 was added for the flowsheet with mutual extraction of 137Cs+90Sr). Stripping of 137Cs was done by ~12 M HNO3, stripping of 90Sr by 0.5 M N2H4 in 2 M HNO3, stripping of nonradioactive Ba by 2 M N2H4 in 4 M HNO3, and reconditioning of the extractant by 3 M HNO3. The stripping of Ba is essential for the process and subsequent modifications. The Ba2+ cation, being better extracted than Sr2+ by PEGs, would otherwise accumulate in the organic phase with loss of extraction efficacy. The first hot cell testing of the chloro-protected bis(dicarbollide) process was done at Gatchina in the early 1980s [146]. The extraction line consisted of 47 mixer-settlers. A two stage flowsheet was used: in the first stage 137Cs was extracted

Copyright © 2004 by Marcel Dekker, Inc.

308

Rais and Grüner

by the H+12- without PEG and in the second stage 90Sr form the raffinate of the first stage was extracted by the H+12-+Slovafol 909 combination. The feed was spiked with a real raffinate of the PUREX process. The extractant returned six times and good efficiencies for the separations of 137Cs and 90Sr from other nuclides as well as good mutual separation of two were reached. Plant testing [147] at Mayak, Russia was performed in the 6-month period of September 1984 to March 1985. Altogether 39 mixer-settlers with a volume capacity of 10 L each were used. The combined flowsheet, with extraction of 137Cs and 90Sr together in the first stage and their consecutive stripping into separate fractions, was used. During the experiments, 15 m3 of model PUREX HLW solution and about 100 m3 of real raffinate of PUREX HLW were used. For experiments with the model solutions 0.5 m3 of the extractant and for experiments with real PUREX solution 3 m3 of extractant were used. This corresponded to about 33 cycles of return of the extractant with real PUREX HLW solution. Generally, stable operation was attained with no changes in performance or hydrodynamic parameters. The efficacy of separation of both isotopes was higher than 98%. From other variants of the process, the possibility of extraction of trivalent lanthanides and actinides with the chloro protected cobalt bis(dicarbollide) extractant must be noted. A variant in which a more concentrated nitrobenzene solution of H+12- is used (~0.3 M) with Slovafol 909 and extraction is done from diluted nitric acid (~0.5 M) was proposed in Ref. 225. This process unit must be inserted in the complete flowsheet after the step in which 137Cs and 90Sr are already separated. A high concentration of the chloro-protected cobalt bis(dicarbollide) is needed in the extraction of trivalent cations and the stripping of 137Cs and 90Sr might be difficult from >0.2 M H+12-. The use of stronger stripping agents, amminocomplexes of Zn, was proposed by NRI and tested in the hot line at Gatchina [147]. Although the obtained data were fully successful, an unexpected obstacle occurred. This was a formation of a solid NH4NO3 deposits on the apparatus surfaces in the hot cell, originating from the contact of evaporated gaseous HNO3 and NH3 in storage vessels. This obstacle could be possibly eliminated by some other means, but no other experiments were done with this system. The process was tested in Russia in a new version [226]. For it 0.3 M H+12- in MNBTF containing 6% of PEG (of undefined type) was used. Fractional stripping provided separate lanthanide and actinide portions. The composition of the stripping solutions was not given in the paper.

B. 1.

New Processes with Chloro-Protected Cobalt Bis(dicarbollide) Process for Extraction of m-Nitrobenzotrifluoride

137

Cs and 90Sr with

The detrimental toxic properties of the mixture of solvents used in the processes described above were partly relieved by a proposed new solvent—m-nitrobenzotrifluoride,

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

309

MNBTF. Moreover, its use does not lead to the formation of corrosive Cl- ions arising from the CCl4 in the previously used compositions and the losses of the solvent into the aqueous effluent are lower than for nitrobenzene. Potentially carcinogenic CCl4 could be avoided, since the density of MNBTF without any additive is sufficiently high (1.467 g/ml, Table 14) to comply with the demands for mixer-settler operation. Besides the new solvent, the process patented in Ref. 210 does not differ from the previously tested process in the Russian-Czech cooperation [8], including the use of H+12- and PEG for extraction and hydrazine for Sr stripping. Still, the new solvent was important for implementing the process on a plant scale operation in Russia. 2.

Process for Extraction of 137Cs and 90Sr with Phenyl Trifluoromethyl Sulfone Diluent

Fluorinated polar solvents, developed originally in Russia and further in the RussianAmerican DOE cooperation, became a new variant of the chloro protected cobalt bis(dicarbollide) process. One variant concerns the first part of a two-stage process in which 137Cs and 90Sr are extracted by the H+12- and PEG combination in fluorinated ethylene glycol. In the second-stage rare earths, technetium and the actinides (especially uranium, plutonium and americium) are extracted from the aqueous phase using a phosphine oxide in a hydrocarbon diluent. The process was patented in the United States in 2001 and in Russia in 2002 [227, 228]. Several reports on the cooperative activities of the Khlopin Radium Institute (RI), St. Petersburg, and Idaho National Engineering and Environmental Laboratory (INEEL) appeared during late 1990s, e.g., Refs. 229–232. These reports did not specify the compositions of extractants and stripping agents; hence they have little informative value. After patenting, full information at last emerged, as given below. A new process for the extraction of 137Cs and 90Sr was developed conjointly in the cooperation between INEEL and RI. The process uses for the extraction a solution of 0.08 M H+12- +0.6 vol% PEG 400 in phenyl trifluoromethyl sulfone (denoted as FS17 in Fig. 6) [233]. The process was tested with simulated evaporated INEEL acidic waste, containing 1.96 M acid, 1.59 M NaNO3, 0.61 M Al, 0.18 M K, 0.006 M Ca, 0.014 M Zr, 0.099 M F-, and 5.64 M NO3- as macrocomponents. Various stripping agents were tested. Particularly low distribution coefficients of Cs and Sr were obtained with the following mixtures: 1.5 M dimethylformamide in 2 M HNO3 (0.037 and 0.049, respectively), 1 M guanidine nitrate in 1 M HNO3 (0.22 and 0.064), and 3 M methylamine in 4.5 M HNO3 (0.053 and 0.033). It was argued that both guanidine and dimethylformamide might not be easily washed from the organic phase. Because methylamine could be easily washed by the acid, this reagent was used in further tests of the study. However, it should be noted that both methylamine and dimethylformamide may form explosive azides during the washing step, which might be prohibitive for their technological use.

Copyright © 2004 by Marcel Dekker, Inc.

310

Rais and Grüner

According to Tables 1 and 3 of Chapter 6, the affinity of CH3NH3+ toward the nitrobenzene phase is comparable to that of K-, so this particular cation is an efficient stripping agent for Cs+ and as seen also for Sr2+. It remains to be resolved whether the stripping proceeds by interchange of CH3NH3+ with Sr2+ cations inside the organic phase complex (the mechanism valid for the much more hydrophobic Slovafol 909 and exchange of doubly protonated hydrazine with Sr2+) or whether all the PEG 400: Sr2+ complex passes into the aqueous phase. If the latter is valid, corresponding changes of the PEG 400 concentration will occur during the Sr2+ stripping, perhaps leading to more difficult process control but not lower economy because of the low price of PEG 400. It may be noted that methylammonium ion as a striping agent was already proposed before in a paper from NRI [234]. 3.

Process for Extraction of 137Cs and 90Sr, Rare Earths, and Actinides with Two Synergists and Phenyl Trifluoromethyl Sulfone Diluent: The UNEX Process

An extensive study was performed within the cooperation between the RI and INEEL, in order to find a system with H+12- and two synergists, which would separate all the title elements from the radioactive waste for “decategorizing” of the waste. This successful research was published in articles [191, 235, 236] reports [237], conference papers (see preceding section), and patents [238, 239]. The process is named UNEX, coined from the words Universal Extraction. In the process, a solution of 0.08 M H + 12 - , 0.02 M diphenyl-N,Ndibutylcarbamoyl- methylene phosphine oxide, DPhDBCMPO, and 0.5 vol% PEG 400 in phenyl trifluoro-methyl sulfone, FS-13 (FS17 in Fig. 5) is used as an extractant. The feed was simulated evaporated INEEL acidic waste, described in the previous section. The title nuclides extracted into the organic phase, were stripped together by a solution of a single composition, namely aqueous 10 g/L diethylene triamine pentaacetic acid and 1 M guanidine carbonate [191, 235, 236]. Protonated guanidinium is a very efficient competing agent for Cs+, according to Tables 1 and 3 of Chapter 6, comparable in its Gibbs energy of transfer to Rb+. Thus, the stripping of Cs+ is probably due to guanidinium+ and the stripping of trivalent lanthanides and actinides is due to the complexing ability of the aqueous complexant at the pH determined by the guanidinium carbonate. This is presumably also the mechanisms of Sr stripping. 4.

Extraction of 137Cs and 90Sr as a Front-End Process Before the PUREX Process

The high radiation stability of H+12- has lead to an audacious project to insert the chloro protected cobalt bis(dicarbollide) process for the separation of 137Cs a 90Sr before the PUREX extraction reprocessing of the irradiated nuclear fuel [240].

Copyright © 2004 by Marcel Dekker, Inc.

Extraction with Metal Bis(dicarbollide) Anions

311

According to the authors of Ref. 240 this concept could have advantages “in essential (in two and more times) reduction of radiation load on tri-n-butylphosphate and a diluent and as a result: reduction of cost of biological protection from ␥-radiation, reduction of explosion and fire risks of PUREX process and its simplification.” Due to the high concentration of uranium in the dissolved fuel, higher concentrations of H+12- and of PEG than in the process for isolation of Cs and 90Sr from PUREX raffinate have to be used, but the process is feasible. For example, with an extractant of the composition of 0.3 M H+ H+12-+6% OP-10 (di-isobutylphenol decaethyleneoxide glycol) in nitrobenzene, the measured distribution ratios were DCs=4.2 and DSr=16.3 for a solution of 250 g/L of U in 4.6 M HNO3. The process was tested under dynamic conditions with 0.3 M H+12-+6% OP-10 in m-nitrobenzotrifluoride, with a wash of the extractant by 3 M HNO3 and consecutive strips with 2 M and 6 M HNO3 (aimed for disposal) and with 4 M HNO3+“organic base” for common strip of 137Cs and 90Sr [240].

C. Plant Scale Operation The main practical result of the application of chloro protected cobalt bis(dicarbollide) extraction chemistry is the plant operation of the process at Mayak, Russia, designed for the fractionation of high level radioactive waste, HLW, as the only plant of the kind in the world. The chloro-protected cobalt bis(dicarbollide) part of the extraction line at Mayak worked in campaign mode, but full information on the work done at the plant has not been fully published from comprehensible reasons. At the end of the 1980s the plant started its operation with the highly toxic and corrosive solvent o-chloronitrobenzene in hexachlorobutadiene, but soon the solvent was changed for m-nitrobenzotrifluoride [18]. The used extractant was 0.06–0.15 M H+12- with 2–3% of PEG in the solvent. During the operations, 65 m3 of nonevaporated and evaporated raffinates of the first extraction the PUREX cycle were treated with the effectiveness of 137Cs and 90Sr uptake of 97 and 99%, respectively. The operation line was constructed in 1995 [18] with total volume of extractors of 2.9 m3 and flowthrough maximal output of 700 L/h. As the extractant, 0.1 M H+12- in 98% of m-nitrobenzotrifluoride and 2% OP-10 was chosen. During only a 3-month campaign in 1996, 210 m3 of highly radioactive defense waste were reprocessed with total β activity of 7 MCi. A practically salt-free concentrate of 137Cs and 90Sr, resulting from the operation, could then be added to the solutions for vitrification. The specific radioactivity of the glass could thus be increased twice, resulting in a proportional economical effect. It was expected that under continuous operation of the line, all of about 4 000 m3 of HLW might be treated in a period of some 5 years [18]. In the Russian report of Romanovskii from 2001, it is stated that by the end of 2000, 600 m3 of HLW was reprocessed and about 25 MCi of 137Cs and 90Sr were recovered [226].

Copyright © 2004 by Marcel Dekker, Inc.

312

Rais and Grüner

VII. ANALYTICAL AND OTHER APPLICATIONS OF EXTRACTION SYSTEMS WITH METAL BIS(DICARBOLLIDE)S A.

Extraction of Metal Cations for Preparative Purposes

1.

Selective Separations

In the systems with cobalt bis(dicarbollide)s and polyethylene glycols, the selectivity of extraction increases as Ca2+precisionD Then GoTo AgainP:, Ds=(KS+LS*P)/Db^2, LogDs=Log(Ds)/Log(10), Cells(11+R, 3)=LogDs, deltaf=59.17*(Log(Db)/ Log(10)-LogKiB), Cells(11+R, 4)=deltaf, Cells(4, 6)=I, Cells(4, 7)=J, logDb=Log(Db)/Log(10), Cells(11+R, 5)=logDb, logP=Log(P)/Log(10), R=R+1, Loop, End Sub Numerical values of the constants KH, LH, Ks (Sr2+), Ls, and ␦P (=delta in the program), and the initial concentrations cH, cB, cp, and cS are placed in the indicated cells into an Excel sheet with a button for starting the macro to which the above macro is assigned (Fig. 1). The calculation fills the data into the columns below the row in which the headings cp, log cp, log Dsr, etc, appear. The value of distribution potential is calculated according to Eq. (25), appearing in the fourth column in Fig. 1. The main characteristics of the systems with maxima are apparent from Fig. 2. As described above, only the value DSr=([Sr2+]or+[SrL2+]or)/[Sr2+]aq) passes through a maximum, whereas all Didion show a monotonic change. Thus, the maximum in the DSr curve is not reflected in the variation of the distribution potential. Some of the curves for different stabilities of M2+ with P are given in Fig. 3. It is seen that a

Copyright © 2004 by Marcel Dekker, Inc.

350

Rais

Figure 1 Layout of the Excel sheet for use with program EXTRIT. The program after filling the constants, concentrations, assumed DB and P values (terms with “def”) and demanded precisions of calculation (terms with “prec.”) starts by the button “calculate.” Besides the results, also the numbers of two iterations are registered (No.it.). These upon terminating the calculation show the results for the last entry of cp.

Figure 2 Calculated distribution of hydrophobic anion B-, bivalent cation Sr2+ (as overall DSr and distribution ratio of individual particle Sr2+), and distribution potential of the system upon extraction from 1 M mineral acid with 0.06 M H+B- in a polar organic solvent and variable concentration of ligand L forming HL+ and SrL2+ particles in the organic phase. The following constants were used: Kex(H+, B-)=10, Kex(Sr2+, 2B-)=5000, Kex(HL+, B-)=1× 107, Kex(Sr2+, 2B-)=1×1012, δL=0.001, log Kex(B-)=8.8, cH=1.06 M, cB=0.06 M, cSr= 0.01 M. Fully dissociated case, calculated by EXTRIT.

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

351

Figure 3 Shapes of maxima for M2+ cations of different extractabilities. The constants employed are shown in the figure. Fully dissociated case, calculated by EXTRIT.

maximum appears only for sufficiently high extraction constants of metal with ligand. For sufficiently high extraction constant LS, the system “mimics” (at the lhs of the two uppermost curves) the classical behavior, namely a plot of log Dsr vs. log cP is a straight line with a slope of 1 as expected for the formation of a 1:1 complex. Notice also that the curves on the right hand side beyond the maximum have the shape of a titration curve, displaying an inflexion point where cP=cB. Other variables tested for convergence were: (1) any number of M+, M2+ and 3+ M ions at macroconcentrations [add respective terms to Eqs. (40) and (41)], (2) vaq/ vor⫽1 [terms with vaq/vor appear in Eqs. (41) and (42)], and (3) formation of aqueous complexes of M with P [respective terms added to the balance of P, changed Eq. (43)].

F. Nonpolar Solvents and Extractions with Mineral Acid Anions For extractions into nonpolar solvents, only the constants with full association in the organic phase do apply. At first sight, and for a long time taken for granted also at the NRI laboratory, the systems with full association are simple and maxima in

Copyright © 2004 by Marcel Dekker, Inc.

352

Rais

the curves ought not to be observed. This, however, is not the case, and it is shown that the existence of maxima is a natural consequence of the mechanism involved in systems where the electrolyte is dissociated at least in the aqueous phase, as also for systems with hydrophilic mineral acid anions. At first, consider the simple case of the determination of selectivity, analogous to the case with full dissociation in the organic phase. In analogy with extraction exchange constant for fully dissociated electrolytes Kexch(Cs+/M+), an extraction exchange constant with full association in the organic phase Kexch(CsB/M+) is defined as (45) It is trivial to show that same relation as Eq. (31) above is valid [40]. The influence of changing the ratio of mineral acid and hydrophobic anions on the extraction was first investigated by Tachimori et al. [41]. Their relationships [41] pertain to the case of full association in the organic phase and are reproduced in a slightly modified form here. For the often encountered case of the extraction of a trivalent cation M3+ in the presence of a neutral ligand L, forming respectively particles (M+, nL, 3B-) and (M3+, nL, 3X-) in the organic phase—where X- and Bare mineral acid and hydrophobic anions—a composite extraction exchange constant of ion pairs with X and B, Kex(M, nL, iB, (3-i)X) corresponding to the equilibrium: (46) applies, where Bor denotes all forms of B in the organic phase and as a rule is represented by cB since the strongly hydrophobic anion B fully extracts into the organic phase in the presence of L. Under the condition where Kex (M, nL, iB, (3i)X), Lor, and are all constant, the equation for DM becomes log DM=ilog [B]or+const

(47)

If plotting the dependence (47) for various series with constant concentrations of X in the series, we can determine for each instance the composition of the organic phase complex, namely the coefficient i. The gradual change of nitrate for the chloro-protected bis(dicarbollide) anion in the organic phase is described in more detail in Chapter 5, Section IV.C. Noninteger values of i found for different aqueous concentrations of nitric acid apparently show a statistical mean of different compositions of the ion pairs formed. Finally, a general relation for the distribution of microamounts of an M3+ cation either in the presence of a hydrophobic anion or a mineral acid anion can be written

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

353

[42]. For the extraction from acid media, there are two extraction constants with bonding Kex(HL, X) and Kex(ML3, 3X), assuming that the hydrogen ion and trivalent cation form 1:1 and 1:3 complexes with L, respectively. These are: Kex(HL, X)=[HLX]or ([H+]aq [L]aq [X-]aq)-1

(48)

and (49) From the last relation, the distribution of microamounts of M is but the distribution of L must be calculated. The latter is not influenced by microamounts of M3+. Taking into account the electroneutrality of the aqueous phase [H+]aq=[X-]aq, and the mass balance relations cL=[HLX]or+[L]or+[L]aq and CH=[HLX]or+[H+]aq and distribution constant of the ligand L, ␦L, we arrive at the cubic relation:

(50)

from which [H+]aq, [L]aq, and DM can be calculated. Thus, the maxima in the curves of DM vs. cL do occur generally, both for hydrophobic or simple mineral acid anions and for the cases of nonpolar solvents as well as for polar ones. The ubiquitous occurrence of maxima observed in extraction of M3+ ions with organophosphorus reagents and malonamides can be explained by Eqs. (49) and (50). The character of the phenomenon is apparent from Fig. 4. The experimental results were obtained with one particular malonamide and extractions of Eu3+ from HCl, HNO3, and HClO 4 media, with B=H+[(1,2-C2B9H11)2-3-Co]-, H+1-, into isopropylbenzene. The experimental curves were compared with those calculated by Eqs. (49) and (50), and the positions of the maxima could be relatively well correlated (with a less precise result for the H+ 1- acid, for which also a larger spread of experimental results was observed). The results are further discussed in Section V.C.

G. Activity Coefficients The extractions are generally performed at concentrations not permitting to neglect the changes of activity coefficients in both the aqueous and organic phases. This subject was tackled by different authors in various forms. The main obstacle is the general uncertainty of how to express the individual ionic activity coefficients especially in the organic phase.

Copyright © 2004 by Marcel Dekker, Inc.

354

Rais

Figure 4 Extraction of Eu3+ from media of mineral acids and H+ [(1,2-C2B9 H11)2-3-Co]-, H+1, by N,N-dimethyl-N’, N’-dibutyl(dodecyloxyethyl) malonamide, DMDBDDEMA, into isopropylbenzene from Ref. 42. Ligand concentration and acid type are shown in the figure. Theoretical curves: dashed lines were calculated using δL=200, and Kex(HL, X), Kex(ML3, 3X) equal to 1×100, 8×104 (HCl); 3×100, 1.4×106 (HNO3); 2×101, 1.6×1010 (HClO4); 1×106; 1×1023 (H+1-). The logs of Kex(ML33+, 3X-) vs. logs Kex(H+, X-) give straight lines with a slope of 2.99 as expected.

The classical approach relies on the use of the extended Debye-Hückel equation [43]. Thus, the ion size parameter a in this equation was taken to be same in water and the organic solvent (1,2-dichloroethane) since “very hydrophilic ions such as Na+ might exist as hydrated ions in water-saturated organic solvent and the influence on a of hydrophobic ions might not be significant” [43]. Kielland’s values of a [44] were used for the simple ions like Na+ and picrate-, whereas for large ions Bu4N+ and BPh4- the a values were assumed to be identical to the crystallographic ion diameters (0.82 and 0.84 nm, respectively) [43]. A sophisticated program for the modeling of the extraction of ion pairs was written and several times improved at the Oak Ridge National Laboratory. In its present form, the program is denoted as SXLSQI [28]. An apparent advantage of the program is its flexibility, it permits the user to test and during the calculation to modify the underlying principles used in the calculation. These are as follows [28]: (1) use of the extended Debye-Hückel equation for low concentrations of salts, (2) use of Pitzer’s treatment of aqueous activity coefficients for higher concentrations of different ions present [45], and (3) use of Hildebrand’s treatment of the activity coefficients in the organic phase, based on the regular solution theory [46]. The published verifications of the model, in spite of its apparent complexity,

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

355

show only the simplest examples of extraction equilibria, such as distribution of one electrolyte between two phases as a dependence on its concentration [28]. Thus, it is difficult to asses the limits of validity of the model. From another paper of the authors from ORNL [47], the impression can be obtained that the maxima in the curves log DM vs. log cL could be explained and treated by the model. On an examination of the proposed fit, it is apparent that the experimental curve with a maximum (e.g., in a concrete example of extraction of cesium by calixarene from nitrate media of various concentrations of NO3–) was explained by the fitting of the experimental curve with chosen (aqueous) activity coefficients on the right hand branch of the curve where DCs decreased [47]. However, for different neutral ligands the maximum occurs at different aqueous acidities, whereas the change of activity coefficients would imply always the same position of the maximum. For the model presented in this review, the maximum is intrinsic to it. Thus, the treatment presented in [47], can be criticized on account of both the used extraction model, and the activity coefficients used at high concentrations of In the cases of distribution of some fully dissociated electrolytes between water and nitrobenzene (Me4N+Pic-, Et4N+Pic-, Ph4As+Cl-), the distribution ratio of the salt was constant over the entire examined range, approximately up to 8×10-2 M initial concentration of the salt [19]. Similarly, the distribution potential of electrolytes between water and nitrobenzene and between water and 1,2dichloroethane [48] was constant in all instances up to nearly 0.1 M concentrations. These observations suggest that upon the distribution in simple systems, the effect of changing of activity coefficients in both phases cancels out or is negligible. A simplified method of treatment of activity coefficients in the systems, used frequently at the NRI laboratory and partly also by Vanura [36], consists in the application of “concentration constants,” CC, in which concentrations are used instead of activities. These, as found in several studies [9, 40], often obey the rule: log CC=log CC0+b(Iaq)0.5

(51)

where CC0 is the value of at zero ionic strength of the aqueous phase Iaq. If the concentration of electrolyte in the organic phases is sufficiently low and extrapolation to Iaq→0 is done, the extrapolated value CC0 may be identified with the standard thermodynamical constant. As shown in the Section IV.C of Chapter 5, in certain cases of extraction into nonpolar solvents, the effect of the change of the activity coefficient of the aqueous HNO3 up to high concentrations of it (ca. 8 M) is either small or some compensation occurs, so that concentrations instead of activities can be used. This finding was based on the fact that the compositions of the ion pair in the organic phase determined from Eq. (47) varied linearly with the concentration but not the H+ activity of the nitric acid in the aqueous phase.

Copyright © 2004 by Marcel Dekker, Inc.

356

H.

Rais

Concluding Remarks on the Mechanism of Extraction of Electrolytes

The peculiar nature of the mechanism of the extraction of electrolytes does not generally conform to the classical log-log analysis applicable to other areas of extraction chemistry. Some intriguing examples of the underlying phenomena were given in this chapter. From them, the existence of maxima of DM, which are often encountered in the extraction science, seems to be the most important. The existence of maxima may be explained also on the basis of competitive reactions in the systems. For the maximum of DM in the dependence on the neutral ligand content cL in the system, at a constant excess concentration of mineral acid in the aqueous phase, the following reactions are to be envisaged. On the ascending part of the curve, the metal cation extraction increases due to the increase of cL. Beyond the maximum, DM decreases because of competition of (MLn+)or for the increasingly formed (HL+) particle. Analogously, in the system with a maximum in the curve of D M vs. aqueous acidity c H at constant concentration of ligand L, two reactions can also be conceived. In the ascending part of curve, the protonation of L applies, forming a HL+ particle, which brings an anion X- into the organic phase according to the electro-neutrality condition. Since the neutral ligand itself in the L form cannot extract Mn+, this process increases the concentration of the extracting species in the system. After reaching the maximum, on the other hand, excess HL+ would cause a decrease of the extraction of Mn+ by competition. However, the above reasoning are not sufficient for explaining the maxima, since the effects would classically lead again into monotonous curves in which two effects simply add. Looking at Fig. 3, the reasoning of this kind may even not be proper (maximum appears only for certain numerical values of constants). The proper ground for the existence of maxima is in the ionic character of the extracted compounds. This is so beyond doubt for fully dissociated electrolytes in the organic phase. However, even if the electrolyte is associated in the organic phase, the ionic particles do exist in the aqueous phase. The systems with dissociation in the aqueous or in both aqueous and organic phases may be compared with solubility relations of salts, which contain also always terms not amenable to log-log analysis. This property is for a pure inorganic salt already contained in the solubility product and in the extractions, either in or The square terms do not exist in the systems conforming to log-log analysis. Attempts to explain the descending part of the curves beyond the maximum by the effect of activity coefficients in the aqueous phase seem to be of less value. For example, we recently studied the extraction of radioactive 137Cs with several crown ethers and calyx-crown compounds. The maxima for extractions from HNO3 (DCs vs. cH at constant cL) were observed in the region of high concentrations of the acid, above 1 M. However, when extracted from HClO4,

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

357

the maxima appeared at acid concentrations around 0.01 M, thus in a very different range of concentrations. The extraction with CO4– anion is higher than with NO3– from reasons discussed in Section V.C.

V.

GIBBS ENERGIES OF TRANSFER AND SOME SEMIEMPIRICAL MODELS

The ionic values of ∆tG° cannot be determined experimentally, but, based on a reasonable extrathermodynamic assumption, they can be evaluated, especially if an independent indirect measure of their validity can be found. During the years large number of Gibbs energy of transfer data for various ions between two solvents accumulated. The reader is directed to a comprehensive review of Kalidas, Hefter, and Marcus for the reassessment of experimental techniques used for the task as well as for the various kinds of extrathermodynamic assumption [49]. The extensively used extrathermodynamic assumptions may suggest that some “correction” to classical thermodynamics, that not only does not define ionic values, but also prohibits dealing with them (unless they appear in sums or differences with balanced charges), is merely introduced. The scientist faces a dilemma: either forget about using single ion thermodynamics, which, in view of all the gained knowledge about ion-solvent interactions does not make sense, or to employ a “new thermodynamics” in which the ionic values have their own meaning. The subject lies far beyond the scope of this review, but the second possibility is considered here as the more useful one [50], and the use of individual ionic thermodynamic values is self-consistent and unambiguous in it. The used experimental techniques leading to the determination of ∆tG° may be divided generally into methods aiming to determine the values involving two pure solvents, ∆tG°(X, aq→or), and those using experimental arrangements in which the two solvents are in contact and equilibrium and ∆tG°(X, aq(or)→or(aq)) values are determined. Since in this review we are primarily interested in the extraction methods, i.e., transfers between mutually saturated solvents, mainly the latter data will be treated here. Further, because of the intensive research with nitrobenzene as a solvent both in extraction as well as in electrochemical studies, the main solvent of interest is nitrobenzene. The choice of this solvent is based not only on the great number of data collected for it, but because the data may be used also for discussing transfers into other even less polar or nonpolar solvents. The data for nitrobenzene as a solvent were gathered by two independent and largely differing techniques, i.e., by extraction methods and by electrochemical methods, mainly cyclic voltammetry at ITIES. If both techniques lead to the same result, the data may be considered as reliable. Selectivity coefficients determined with ion selective membranes serve for further establishment of the consistency of the data.

Copyright © 2004 by Marcel Dekker, Inc.

358

Rais

The extraction methods of the determination of ∆tG°(X, aq(or)→or(aq)) have already been discussed above. They rely on the determination of the proper extraction constants and the division into ionic contributions that was made almost exclusively with the use of the TATB assumption. Electrochemical methods deserve a short introduction as given below.

A.

Electrochemical Methods of Determination of ∆tG°(X, aq(or)→or(aq))

With the development of the studies at ITIES (interface of two immiscible electrolyte solutions), a number of the ionic ∆tG°(X, aq(or)→or(aq)) values was determined by electrochemical methods. In the past, several studies were performed in which the values were determined of a static electrochemical cell, with beforehand mutually saturated aqueous and organic phases and an appropriate salt bridge between the compositionally different phases [48]. In more modern studies an imposed outer potential is usually applied to the cell and the half-wave potential of an ion X of cyclic voltammograms or polarograms is measured. Reversibility of the wave is a prerequisite for obtaining the correct value of ∆tG°(X, aq(or)→or(aq)). Even if in recent studies employing ITIES the two phases are not as a rule preliminary equilibrated, the systems are considered to be in equilibrium and the measured values are ∆tG° (X, aq(or)→or(aq)) and not ∆tG° (X, aq→or). This is because in the very thin interfacial layer of ITIES the mutual saturation of the two solvents proceeds very quickly and completely, especially if the necessary check of reversibility is done [51]. The electrochemical cell usually consists from two liquid phases brought into contact, one organic and one aqueous with appropriate electrolytes in them (a “supporting electrolyte” in the organic phase) in order to ensure small resistivity of the cell. Two reference and two working electrodes are used, with Luggin’s capillaries for the reference electrodes in order to have their openings near the interface of the liquid phases. The organic reference electrode is made of an aqueous solution near the metal electrode and a salt bridge which usually also ensures the elimination of the potential drop in the reference electrode or permits its control. The experimental techniques advanced largely after Samec et al. [52] introduced the four-electrode potentiostat with IR drop compensation by means of a positive feedback loop and the method of choice is cyclic voltammetry. Nowadays, sophisticated techniques are used and for the purpose of this review are described briefly in notes to the Tables 1–3. The half wave potential determined by cyclic voltammetry relates to the standard distribution potential as follows [53]: (52)

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

359

where Dor(X), Daq(X), ␥or(X), ␥aq, ␥aq(X) are the diffusion and activity coefficients of the ion X in the organic and aqueous phases. (Note: In electrochemical techniques D denotes a diffusion or mobility coefficient, not a distribution ratio as in extraction.) The factor A is connected to the actual geometry of the arrangement, e.g., A=[(4d/␲r)+1] for a microinterface located in an organic solvent-filled microhole of radius r and depth d [53], and the other symbols have their usual meaning. The diffusion coefficients are not always known and two methods are used for their estimation: (1) the Nernst-Einstein relation gives |zi| FDi=RT␭i, where ␭t is limiting ionic conductivity of an ion i, and (2) Walden’s rule states that ␩␣D␣=␩ßDß of two solvents ␣ and ß, where ␩ is the viscosity [53]. For two solvents of similar structure as are, e.g, nitrobenzene and o-nitrophenyl octyl ether, methods (1) and (2) usually lead to identical result, but exceptions were also noted [53]. In the presence of a complexing agent an effect, called “synergy” in extraction, occurs, i.e., a shift of termed “facilitated transfer” in electrochemistry. The stability of the complex in the organic phase can be calculated from the shift of Of course, the stability constants of the complex ought to be identical with that obtained from extraction experiments. The selectivity constants may be determined also from potentiometric measurements when an organic solution of a suitable electrolyte is used as an ion selective electrode, ISE. The measurements in this case provide upon extrapolation of the measured apparent selectivity coefficients to zero ionic strength a “theoretical selectivity coefficient” This quantity is independent of the ion-exchange site, its concentration in the membrane or of the activities of the ions in the test solution, but depends only on the properties of the membrane solvent used [54]. The is related to the extraction exchange constant Kexch(A/B)=Kex(A)/Kex(B), where Kex(A) and Kex(B) are individual extraction constants of the ions A and B, as reported in Refs. 54–56: (53) where DL is a mobility of free ligand in the ISE phase. The relation is often simplified as an approximation to [54, 56]: (54) having practical significance. In fact, for much more complicated conditions, i.e., with the ISE solvent p-nitrocumene embedded into a PVC membrane, Scholer [56] found a strikingly linear correlation of for nine alkali metal and tetraalkylammonium cations with the K ex(A)/K ex(B) measured in waternitrobenzene [19] with a correlation coefficient R2=0.9958. One reason for the apparent success of such a correlation is that PVC in the membrane behaves as an

Copyright © 2004 by Marcel Dekker, Inc.

360

Rais

inert diluent, thus not influencing the selectivity of the system; see Ref. 57 for cation loading experiments and Ref. 58 for electrochemical measurements with 0nitrophenyl octyl ether+PVC membranes. Thus values may well serve as an auxiliary set of values for ⌬tG°(X, aq(or)→ or(aq)).

B.

Ionic ∆tG°(X, aq(or)→ or(aq)) Values for the Water-Nitrobenzene System

A large data set for the individual ∆tG°(X, aq(or)→or(aq)) is now available. The initial data on the transfer of 22 ions obtained by extraction [19] were thoroughly checked and the set was enlarged both by extraction and electrochemical methods. These data are collected in the well-maintained web database of Girault [20] providing values for seven polar solvents and for nitrobenzene and containing almost 100 entries (sometimes with multiple entries for one ion). Another comprehensive source of recommended values is that of Osakai [21, 22] supplying values for 37 different ions. The data for nitrobenzene as a solvent, due to Senkyr et al. [59, 60], were recalculated into the form of individual transfer values based on the TATB assumption by Makrlik [54, 61]. There are 41 entries for cations [54] and 46 entries for anions [61]. A further set for the purpose of comparison is that of Scholer for p-nitrocumene [56], containing 54 different cation entries. Several important sets of data had to be omitted here for the sake of brevity. These are, e.g., seven substituted ethylenediamine derivatives studied by the electrochemical method by Baruzzi and Wendt [62] or ten aromatic amine derivatives studied by the same method by Marecek and Samec [63]. Takeda et al. studied the transfers of amino acids in the system [64]. The width of contemporary determinations of ion transfers data may be illustrated by a recent publication [65], in which the transfers of alkali metal cations between water and nitrobenzene with monoaza-18-crown-6 modified fullerene were studied. Smaller sets of anionic data for various cobalt bis(dicarbollide) derivatives published in [66, 75] are also important. This review is limited to reporting the most important data for small inorganic ions, tetraalkylammonium cations, some complex ions, and the most important hydrophobic anions. When several data existed for one ion, the preferred values are printed in the tables in bold characters. The data are organized into the three following tables: Table 1 in which the data for small inorganic cations are collected, Table 2 for inorganic and dicarbollide anions, and Table 3 for tetraalkylammonium cations plus some cations which are important in extraction chemistry. The Gibbs energies of transfer for alkali metal cations were determined from selectivities of different extraction systems and extraction data are thus preferred. All electrochemical data for the alkali metal cations series are slightly less positive, but Osakai in his recommended values [21, 22] also preferred the extraction values. The opposite is true for H+, and because of the stability of cobalt bis(dicarbollide), we prefer the second entry for H+, which is based on the extraction data with this

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

361

Table 1 Transfer Gibbs Energies of Simple Cations Between Mutually Saturated Phases of Water and Nitrobenzene (NB), ∆tG° (X, aq(NB)→NB(aq)), kJ/mol, 25°C

Data obtained from extraction experiments in which the selectivity of transfer of two ions was determined and the absolute value was obtained on the basis of the TATB assumption (the same extractability of Ph4As+ and BPh4– from water into nitrobenzene). b Recommended values from extraction and electrochemical experiments by Osakai [21, 22]. c Electrochemical database of Girault (containing also extraction results) [20], mainly with the TATB assumption, but see [20] for details; in the column “Electrochemical data” also other electrochemical data are given. d Data calculated by Makrlik [54] on the basis of experimental data of Sustacek and Senkyr [59]. Nitrobenzene contained 10-4 M solution of tetrakis(4-fluorophenyl)borate with a suitable cation and measurement was done with vigorous stirring of the phases. e From [36]. f From selectivity constants Kexch(Cs+/M+) determined by extraction, the reported values being mean values from several other publications [19]. g Two different values of Kexch(Cs+/H+) were determined; the value 33.7 was obtained with cobalt bis(dicarbollide) [9], which permitted better extrapolation to zero aqueous ionic strength and this value is preferred. h From [67], based on measurements on a liquid/liquid microinterface with no electrolyte in the aqueous phase. i Extraction results from [68]. J From cyclic voltammetry [69]. k From [70]. l From solubility measurements, the data referring to pure solvents are consequently put into brackets [71], m Cyclic voltammetry [72]. n AC cyclic voltammetry [73]. o Personal communication of Hundhammer and T.Zerihun to Wilke quoted in [67]. p Extraction data from [74]. qElectrochemical data [75].r From [76], the value for Cs+ was evaluated by the authors of the paper as the mean best value for this cation with a standard deviation of ±0.2 kJ/mol. s From [77]. t Extraction results [33]. u Cyclic voltammetry at the microinterface of a hole of 12 µm in a polymer foil [78]. v Extraction data from [79]. w Extraction data from [80] [i]. a

Copyright © 2004 by Marcel Dekker, Inc.

362

Rais

Table 2 Transfer Gibbs Energies of Anions Between Mutually Saturated Phases of Water and Nitrobenzene (NB), ∆tG°(X, aq(NB)→NB(aq)), kJ/mol, 25°C*a

*Cobalt bis(dicarbollide) anions, numbered as in Ref. 10. a,b,c See Table 1. d Data calculated by Makrlik [54] on the basis of experimental data of Senkyr and Kouril [60]. Nitrobenzene contained 1–2×10-4 M crystal violet cation with the suitable anion and measurement was done with stirring of the phases. e Extraction data from [66]. f Cyclic voltammetry at three phase junction [81]. g Extraction study [19], all data also collected in [20] with exception of the value for I–3 the value in the present table being correct. h Flow-injection amperometry with CIO4– as an internal standard [82]. i From solubility measurements, the data refering to pure solvents are consequently put into brackets [71]. j Cyclic voltammetry [72]. k AC cyclic voltammetry [73]. l Square-wave voltammetry with internal standard CIO4– [83]. m Flow-injection amperometry with CIO4– as an internal standard [82]. n Cyclic voltammetry [72]. o from [82]. p From [72]. q From [82]. r From [72]. s From [72]. t From AC and DC cyclic voltammetry [83]. u From [73]. v Cyclic voltammetry in [84], w The anion of phosphomolybdic acid behaves in acidic media upon transfer as an relatively hydrophobic univalent anion PMo–, see [4] for details. The determined extraction constants of H+ PMo- for various acidities was extrapolated to zero aqueous ionic strength and the value in table was calculated using the individual extraction constant of hydrogen ion. Osakai in his papers gives several values for 3-to 6-charged heteropolyacid anions, see original papers [21, 22]. x From [84]. y Cyclic voltammetry at a microinterface of a hole of 12 µm in a polymer foil [78]. z Tentative value for OH- ion from [85], ref. 11 in the paper to unpublished results of the author. 2,6-DNPh- and 2,4-DNPh- denote 2,6dinitrophenolate and 2,4-dinitrophenolate, respectively, ␣-hexylate is 2,4-dinitro-N-picryl-1naphtylaminate, and the data for substituted cobalt bis(dicarbollide) anions are given in the last column of the table with abbreviations given in the respective tables found in Chapter 5 of this volume [10].

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

363

anion. The y=∆tG°(X, aq(NB)→NB(aq)) values of simple inorganic M+, M2+, and M3+ cations from Table 1 can be correlated with the respective values from [94] by the equation: y=-6×10-20x6–2×10-15x5–3×10-11x4–1×10-7x3 -0.0002x2–0.2344x-29.349

(55)

with a correlation coefficient R2=0.9962 (35 ions from the Table 1, with data for Ag+ and UO22+ omitted since they are outliers to the given correlation). Alternatively, the data can be correlated as separate straight lines when plotted against for the series of M+, M2+, and M3- cations, the slope of these lines decreasing considerably in this series. The lower selectivities of extraction for M2+, and especially M3+ are apparently connected with the extensive hydrations of the cations in the organic phase; see data in Section V.C.5. The ⌬ tG°(X, aq (or)→or (aq) ) data for transfer of Cs + ion from water to nitrobenzene+CCl4 mixtures are important for the previously studied transfers into the mixtures of nitrobenzene with non-polar solvents. The data obtained by cyclic voltammetry are as follows (vol% of nitrobenzene in the mixture/∆tG° in kJ/mol): 100/13.2, 95/14.3, 90/15.3, 85/16.3, 80/16.8, 70/17.9, and 60/19.4 [95]. The potentiometric selectivity coefficients yield results for small hydrophilic anions different from extraction and other electrochemical data, but this is not important, since this set is only an auxiliary set. The agreement for extraction and electrochemical data of anions is acceptable with exception of Cl- ion. The gap of nearly 10 kJ/mol for Cl- anion cannot be reasonably explained. Both values are apparently reliable, but the extraction value is given more weight in view of the correlation given in Section V.C.5. The value for Br- was not checked by extraction, recommended value by Osakai is quoted. The behavior of was determined only at the NRI laboratory and was not confirmed by others. The tetraalkylammonium ions give practically same results in all columns with exception of the data for Et4N+. However, its value was twice determined by Osakai [21, 22, 87], hence his newer and by him recommended value [21, 22] is deemed to be more reliable. The data for complexes of alkali metal cations with two crown ethers, although not substantiated by other measurements, seem to be correct. The ions become substantially more hydrophobic if enveloped by an organic hydrophobic moiety.

C. Physical Chemistry of Transfer: Empirical Models Several theoretical approaches have been used in order to interpret the Gibbs energies of transfer between two solvents. While the reader may find the respective theories in the literature, e.g., Refs. 22, 27, 28, the aim of this review is different. We try to find some simple empirical relations and indices that can serve as a practical tool for predicting the properties of the systems. Normalized functions may be used for checking whether a particular data point does or does not conform

Copyright © 2004 by Marcel Dekker, Inc.

364

Table 3 Ionic Values of Transfer Gibbs Energies of Alkylammonium and Some Ions Used in Extractions from Water to Nitrobenzene ∆tG°(X, aq(NB)→nitrobenzene(aq)) Between Mutually Saturated Phases, kJ/mol, 25 °C

Rais

Copyright © 2004 by Marcel Dekker, Inc.

365

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

Symmetrical compounds noted by s after the compound number. a From [86], extrapolated Kexch(Cs+/M+) constants to zero ionic strength, 0.06 M H+ 1- in nitrobenzene. b From [19]. c Data calculated by Makrlik [54] on the basis of experimental data of Sustacek and Senkyr [59]. Nitrobenzene contained 10-4 M solution of tetrakis(4-fluorophenyl) borate with suitable cation and measurement was done at vigorous stirring of the phases. d From [87], by cyclic voltammetry. e Recommended values from extraction and electrochemical experiments by Osakai [21, 22]. f Recalculated from standard distribution potentials determined from AC polarography at water-NB interface from [88] as referred in [89]. g Database of Girault [20]. h From [84]. i From solubility measurements, the data refer to pure solvents and are consequently put into brackets [71]. From [72]. k From solubility data [90]. L Cyclic voltammetry at a microinterface of a hole of 12 µm in a polymer foil [78]. m From [91]. n From [92]. o Approximate value from aqueous medium 1 M HNO3, calculated by EXTRIT. The data were obtained for nitrobenzene+CCl4 mixture (60 vol.%:40%) and under justified assumption of same selectivity for this mixture and pure nitrobenzene are used here [93].

366

Rais

to the trend displayed by the majority of other data. The section deals with the transfers of small hydrophilic ions, large hydrophobic ions, data for other polar solvents, the correlation of solubility and extraction, hydration of ions in the organic phase, role of dissolved solvent on the selectivity, and general affinity scale for organic solvents. 1.

of Inorganic Ions in the Water-Nitrobenzene System

The subject of ∆tG° values between pure solvents was recently discussed by Rais and Okada [96]. It was shown in the paper that the choice of the entered experimental data is important. If the data of Gritzner, based on the measurements of the potential of reduction of alkali metal amalgams and performed at one laboratory [26], were used, a relatively simple picture of solvation could be conceived. The values of ∆tG° in the chosen coordinates were well expressed by their linear dependences on the Gibbs energies of hydration of the respective ions. The enigmatic, but often noted property of such systems, i.e., that the behavior of the first complex between the cation and solvent molecule frequently mirrors that of the overall ∆tG°, could be tentatively explained by the formation of energetically quantized clusters of the ion with the first four solvent molecules [96]. An analogous behavior is displayed also for the mutually saturated solvents. The selectivities of the transfers of alkali metal cations from water obtained with 137Cs spiked solutions according to the method described in the Sections IV.C and IV.F. [40] were determined. These selectivities: Kexch(Cs+/M+) for polar solvents or Kexch(CsB/ M+) for essentially non-polar solvents, supplemented with older ones and expressed as relative values, are shown in Fig. 5 and Table 4, respectively, for seven new solvents or solvent mixtures. In Table 4, the data for transfers between pure solvents, recalculated from those by Gritzner [26] to the values with water as a reference solvent, are also included. The reasonableness of the both sets of values is corroborated by the fairly straight line correlations of DtG° or with The data for Na+, K+, + + + Rb , and Cs correlate as straight lines with y=[∆tG°(M )-∆tG°(Cs+)], R= correlation coefficient: dry nitromethane y=-0.229x-58.076, R2=0.996; o-nitrophenylether saturated with water y=-0.1098x-27.529, R2=0.9994; nitromethane saturated with water y=-0.1129x-27.997, R2=0.9972; nitrobenzene saturated with water y=-0.1648x-41.124, R2=0.999. The data of standard molar Gibbs energies of hydration are from Ref. 94. The hydration effects in the organic phase mostly concern the Li+ ion and its transfers are relatively increased into water saturated solvents compared to pure solvents, see the dependences for pure and water saturated nitromethane. As concerns the steepness and sign of the dependence, a general rule that the slope of the straight line correlations of with is primarily determined by the donor number, DN, of the solvent [96] is obeyed. This is exemplified by the

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

367

Figure 5 Transfers of alkali metal cations relative to Cs+ for various water/mutually saturated organic phase extraction systems (except “dry NM”) at 25 °C (From Refs. 40 and 97; data for nitromethane are from Ref. 99.)

most negative slope at Fig. 5 displayed by nitromethane (lowest DN) and the most positive by tri-n-butylphosphate (highest DN). The points for four alkali metal cations give a straight line with R2>0.996 also for a mixture of 70% vol of NB+30% CCl4, thus justifying the choice of as a normalizing function. For PC, DOS and DOA, the data point for Rb+ lies outside the correlation, but the reasons for this particular deviation remain unclear. Consequently, the recommended values for transfers of small cations and anions from water to nitrobenzene from the Tables 1 and 2 are plotted against their respective as shown in Fig. 6. The values for simple inorganic cations and anions fall on a single straight line. The only exception is Li+, probably because of its excessive hydration. This finding says that the standard molar Gibbs energies of transfer of small cations and anions from water to nitrobenzene are same for a given value of irrespective of ionic charge. Thus, for the reverse transfer of both cations and anions

Copyright © 2004 by Marcel Dekker, Inc.

368

Rais

Table 4 Ion Transfer Values of Alkali Metal Cations: ∆tG°(M+, aq→or) Between Pure Water and Pure Solventsa and ∆tG°(M+, aq(or)→or(aq), Csrel+) Between Mutually Saturated Solventsb, kJ/mol, 25°C

Abbreviation of the solvents: methanol MeOH, ethanol EtOH, 1-propanol PrOH, 1-butanol-BuOH, 1,2ethanediol EG, propylene carbonate PC, trimethylphosphate TMP, N-methylformamide MF, N,Ndimethylformamide DMF, N-methyl-2-pyrrolidinone NMP, hexamethylphosphoric triamide HMP, dimethyl sulfoxide DMSO, acetonitrile AN, propanenitrile PRN, benzonitrile BN, pyridine PY, pyrrole PL, N,N-dimethylthioformamide DMTF, N-methyl-2-thiopyrrolidinone NMTP, nitromethane NM, nitrobenzene NB, o-nitrophenylether NPOE, 1,2-dichloroethane 1,2DClE, Dioctylsebacate DOS, Dioctyladipate DOA, Tri-n-butylphoshate TBP, Toluene TO. a Recalculated from the tabulated values ∆tG° for acetonitrile→solvent transfer from [26]. b Results of new studies, ∆tG°(M+, Csrel+) [40, 97, 98]. See Sections IV.C and IV.F for experimental conditions for “ms” solvents. In some cases (1,2-dichloroethane), a lower concentration of a more hydrophobic cobalt bis(dicarbollide) derivative had to be used because of solubility problems [40, 98]. c p=pure solvents, ms=mutually saturated solvents, ass=associated in the solvent. d From [94]. e From [99].

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

369

Figure 6 Correlation of cationic and anionic data for transfer from water to nitrobenzene. ∆tG°(X, aq(or)→or(aq)) values are the recommended ones in Tables 1 and 2 (X denotes the – value proposed by Osakai for transfer of Cl ). Standard molar Gibbs energies of hydration are from Ref. 94. Hydration numbers of ions in nitrobenzene are those from Table 5 and Refs. 22 and 111.

from nitrobenzene into water the hydration in the aqueous is the main driving force. This reasoning is corroborated by the independence on charge of the hydration numbers in nitrobenzene. A self-consistent picture emerges from Fig. 6: (1) individual ∆tG°(X, aq(or)→or(aq)) values are to be considered as real and appropriate into its numbers, and (2) the values and the method of division of total ionic values [94] are validated. Although in the present paper the data for small inorganic anions and all available solvents were not analyzed, the situation should be similar to that of the cations as concerns the normalizing function. This is based on the finding that the gaseous cluster Gibbs energies of formation for the n=1 to 6 clusters for alkali metal cations and halide anions yield straight lines when plotted against the pertinent hydration energies [100]. The total hydration energetic is expressed already in the simplest 1:1 ion: molecule clusters and all subsequent ones as well, and provides a plausible explanation of the applicability of the chosen normalizing function. The extraction technique used for the determination did not distinguish between dissociation and association in the organic phase. In low polarity DOS, DOA, and toluene the predominance of ion pairs is proved by conductivity measurements. In these cases, the selectivities given in Table 4 pertain to the exchange of the aqueous ion with the organic associate. However, the similarity of the absolute values and

Copyright © 2004 by Marcel Dekker, Inc.

370

Rais

shapes of the curves in Fig. 5 indicate that ion pairing in the organic phase should not be of importance as far as the selectivity of the system is concerned. Very high selectivities, which were found for toluene as the solvent, appeared also in our previous studies with mixtures of nitrobenzene with some nonpolar diluent. The data for a mixture of 30% of nitrobenzene with 70% of CCl4 are given in Fig. 5 and Table 4. Another conspicuous example of a dramatic increase of selectivity of extraction of strontium is provided by the data plotted in Fig. 7 [101]. The same effect occurred for extraction of Cs+ in the presence of Rb+ and the selectivity apparently increases in all the row of alkali metal cations. A method for how to attain an increased separation factor of the Cs+/Rb+ pair in extraction with cobalt bis(dicarbollide)s was described in Ref. 102. Whereas for the system water/ nitrobenzene the separation factor DCs/DRb is 5, for a mixture of 30 vol% of nitrobenzene+ 70% of benzene DCs/DRb was 16. A similar effect occurred when carbon tetrachloride replaced benzene [102]. The effect of the addition of a nonpolar solvent can be explained as follows. The basicities of nonpolar solvents such as toluene or benzene are not known (no DN data) but presumably are low. The high selectivity may thus be due to a decrease of the basicity of the solvent upon addition of the nonpolar solvents. On the other hand, electrostatic interactions connected with ion-dipole terms decrease with decreasing ε, and this may also be a leading factor. The data plotted in Figs. 5 and 7 express the upper limit of selectivity attainable in the practical systems, since for still lower ε, the solvation of the hydrophobic anion will also drop, and the extractant will pass into the aqueous phase. This is actually the case according to our measurements.

Figure 7 Extraction of Sr2+ from 0.5 M HNO3 in the aqueous phase by 0.001 M bromoprotected cobalt bis(dicarbollide), H+[(l,2-C2B9H8Br3)2–3–Co]-, in the presence of PEG 400 for various compositions of the organic phase. Curve 1:10 vol% of nitrobenzene (NB)+ 90% benzene, 2:20% NB, 3:30% NB, 4:50% NB, and 5:100% NB.

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes 2.

371

of Tetraalkylammonium Ions in the Water-Nitrobenzene System

A large amount of attention was paid to the transfer of voluminous organic ions and was treated by several empirical approaches besides other well-known methods [27]. For large nonhydrated ions of this type it may be reasonably supposed that ∆tG°= ∆tG°(X, aq(or)→or(aq)) and this shortened notation is used here. One method uses Uhlig’s formula for expressing the nonelectrostatic part of ∆tG°, ∆tG° (ne); see e.g, Osakai’s theory [21, 22]. In the expression, ri is the radius of the ion and σo,aq is the interfacial tension at water— organic phase interface. The ∆tG° values for transfer from water to nitrobenzene show a constant increment of -2.5 kJ/mol per methylene group [87]. The total ∆tG° is then given by the nonelectrostatic contribution, or, at least, the relative values in a series are independent of the charge of the ion or any electrostatic interaction. Thus, we can expect some simple function of the radius of an ion for the magnitudes of ∆tG°. The van der Waals radii (Me4N+ 280, Et4N+ 337, Pr4N+ 379, Bu4N+ 413, Ph4As+ 425 pm [49], Ph4B- 421 pm [22]) are most often used for the voluminous ions. The difference of ∆tG° for the pair Bu4N+/Ph4As+ is contrary to expectation if the van der Waals radii are used. Before assigning a special property to this pair, we should check whether some other scale of radii might be more appropriate. One such set, closer also to other proposals of crystallographic radii of the ions under question, is that of Krumgalz [103] (216, 267, 335, 385, and 428 pm, respectively, for the cations in the previous paragraph), the difference of ∆tG° for the Bu4N+/Ph4As+ pair falling in line with the tetraalkylammonium cations. A further set of ion radii can be independently calculated, based on the so called “covalent surface area” empirical method [50]. The surfaces of all atoms contained in the ion (using tabulated covalent radii) are summed and the effective radius rCSA is back recalculated. Although the method is based on employing additive contributions and avoiding atom overlaps that appear in the van der Waals radii, its agreement with Krumgalz’s radii is relatively good (200, 267, 321, 367, and 441 pm, respectively). Because of its simplicity, this method may be used for the correlation of the ∆tG° of tetraalkylammonium and related ions. However, the importance of using a particular set of radii or deciding whether a plot against r or r2 is to be used should not be overestimated. In fact, when one tries to express the ∆tG° data contained in the set [87] with the above choices, the best correlation coefficient of the straight line is obtained when ∆tG° values are plotted against the number of carbon atoms in the ion. A similar linear correlation for transfers of alkylammonium ions from water into NPOE solvent (embedded in a PVC membrane) on the number of carbon atoms was reported recently in Ref. 104. Because of noted constant increment per CH2 group, the values may be as well correlated with the partition constants ␦BS (aq→n-octanol). A relatively good similar correlation was obtained for cobalt bis(dicarbollide) anions in Ref. 105.

Copyright © 2004 by Marcel Dekker, Inc.

372

Rais

For more complicated ions that are not the subject of this review, the correlation of their ∆tG° values based on topological indices, describing in more detail the size and also space orientations of the substituents, was used in Ref. 106. Although this method should be preferred for more complicated ions, the reported agreement for simple tetraalkylammonium ions from Ref. 87 was poorer than when a constant increment per alkyl group was used. 3. Other Polar Solvents In conclusion to the data on ion transfer for the water—nitrobenzene system it can be said that the two methods, extraction and electrochemical, although completely different in concept, furnish very near results, thus mutually corroborating one another. As concerns the transfer to other solvents, the data, not as extensive as for nitrobenzene, cannot be treated here in detail. It is useful to note that Wilke [53] correlated transfer data from water into o-nitrophenyl octyl ether with transfers from water into three other organic solvents: nitrobenzene, 1,2-dichlorethane and o-dichlorbenzene (Fig 8). The linearity for small and large ions is surprising,

Figure 8 Correlation of transfer data among o-nitrophenyloctylether and nitrobenzene, odichlorbenzene, and 1,2-dichloroethane solvents according to Ref. 53. (Reprinted from S. Wilke, T.Zerihun, J. Electroanal. Chem. 515:52, 2001, with permission from Elsevier.) The abbreviations are self-explanatory; DS denotes n-dodecyl sulphate anion. The correlations in [53] equals to ∆tG° by the authors are written in the upper left corner of the figure. here, in ␦J/mol units.

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

373

and could not be explained by the Born equation or other electrostatic theories [53]. This correlation serves here for three purposes: (1) Instead of full tables, it shows the data for transfer to polar solvents other than nitrobenzene, (2) the linearity of the ∆tG° on for simple cations and anions is corroborated by the figure, and (3) a tentative scale of anion hydrophobicity based on the data for transfers in the water—nitrobenzene system can be constructed also for transfers into relatively nonpolar solvents (represented in Fig. 8 by o-dichlorobenzene with ε=10.36). 4. Extraction and Solubility An empirical rule, that salts of cesium with anions like dipicrylaminate or tetraphenylborate form insoluble precipitates in water are also extracted efficiently into nitro-benzene, served for finding new extraction systems for this cation [1, 2]. The analytical method for the determination of polyethylene glycols (PEGs) by their precipitation with tetraphenylborate in the presence of Ba2+ [107] suggested the synergetic effect displayed by PEGs for the extraction of alkaline earth cations [108]. In a reverse manner, the precipitation of protonized crown ethers and some phosphororganic reagents with phosphomolybdate or other hydrophobic anions were devised. The dependence of the logarithm of distribution ratio of potassium salts of several hydrophobic anions between water and nitrobenzene on the logarithm of their aqueous solubilities was reported by Iwachido [109]. For 16 anions a straight-line correlation with a slope of -1.11 resulted with R2=0.8713 which is significant, considering that a range of 5 orders of magnitude of the distribution ratios and solubilities was covered. Although the crystal energies of the precipitates formed by the hydrophobic anions are not known, the effect is tentatively ascribed to a loss of affinity toward water. The total standard molar Gibbs energies of water—nitrobenzene transfers of the salts under consideration are slightly to strongly negative (e.g., -24 kJ/mol for Cs+ DPA- compared with -1.1 for the water soluble Li+ DPA-). Hence, the low solubilities are explained by an assumed positive or nearly positive Gibbs hydration energy of the ion combination, thus not enabling its dissolution in water. 5. Hydration of Ionic Species in the Organic Phase From the outset of the studies, the number of coextracted water molecules with a particular ion in nitrobenzene was studied. For this purpose, usually a Karl Fisher titration of the water in the organic extracts is done. The water content of the organic phase, c(H2O)org, generally increases linearly with the concentration of the electrolyte according to the relation where is the water content of the solvent with no added electrolyte. If an

Copyright © 2004 by Marcel Dekker, Inc.

374

Rais

electrolyte N+Y- can be found for which throughout its entire concentration range , then nhyd(N+)=nhyd(Y-)=0. For any other salt M+Y-, the hydration is due to M+ only and the ionic values can be determined. Full dissociation of N+Y- and M+Y- in the organic phase is assumed or incomplete dissociation must be corrected for. Hydration numbers, nhyd of several examined cations with use of the cobalt bis(dicarbollide) anion 1- were determined by the Karl Fisher method in [110]. These were based on the experimentally verified n hyd =0 for the cobalt bis(dicarbollide) anion 1- anion and are nhyd(Li+)=6.5, 6.0, 5.8±0.6; nhyd(H3O+)= 4.5; nhyd(Na+)=3.9, 3.8, 3.7±0.2; nhyd(K+)=1.5, 1.0, 1.4±0.1; nhyd(Rb+)=0.8, 0.7, 0.77±0.06; nhyd(Cs+)= 0.5, 0.4, 0.4±0.3; and nhyd(Et4N+)=0.0, 0.0. The first entries in this list are from Ref. 110, the second entries are the revised values of Osakai [111], and the third entries are the mean values from Ref. 110. All tetraalkylammonium cations and the studied organic anions are nonhydrated in the organic phase, whereas the following nhyd were reported [22, 111]: for Cl- 4.0, for Br- 2.1, for I- 0.9, for SCN- 1.1, and for NO3– 1.7. A trend of more positive ∆tG° values with increasing nhyd is observed for alkali metal cations, and most anions (except SCN- and I-). As an auxiliary graph, the dependence of ∆tG° on nH gave a smooth curve for nine univalent cations and anions and even for uni-, bi- and trivalent cations [80], showing the importance of hydration in the organic phase for the magnitude of ∆tG°. The hydration numbers of Ca2+ and Ba2+ in nitrobenzene are nhyd=14 and 11, respectively [22], the latter value being in agreement with a previously published one (11.5±1) [79]. For trivalent Ce3+, nhyd=16.2±2 was measured in [80]. For complexed alkali metal cations, as expected, the value of nH decreases. Thus, e.g., for complexes of M+ with dibenzo-18-crown-6, DB-18-C-6, the hydration numbers are shown in Table 5. These numbers show that the largest decrease of hydration between a bare and a complexed ion is for Li+, whereas the hydration numbers of Cs+ and Rb+ are low and nearly the same for uncomplexed and complexed ions. Furthermore, due to relative independence of nhyd on the length of PEG complexants, it can be inferred that the coordination sites of alkali metal cations are saturated to the same degree for all alkali metal cations by these compounds. For primary, secondary, and tertiary ammonium ions in nitrobenzene, the nhyd values are 1.64, 1.04, and 0.66, respectively, with the water molecules probably directly bound to the central nitrogen atom [114]. The effect of dissolved the water in nitrobenzene and some doubts concerning the physical sense of the hydration numbers determined by Karl Fisher titration were discussed in a previous review [28]. Hydration of Cl-, Br-, I-, NO3–, NO4–, and SCN- anions in deuterated nitrobenzene saturated with water was studied recently, using 1H NMR spectrometry [115]. The study indicated that the hydration proceeds in a stepwise manner in nitrobenzene, i.e. according to the reaction X-(H2O)m-1+H2O=X-(H2O)m (m=1,2, 3,…). Thus, nonintegral values of the hydration numbers in nitrobenzene found in various papers

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

375

Table 5 Hydration Numbers nhyd of Bare and Complexed Cations in Nitrobenzene Saturated with Water

Mean values from [110]. From[112]. c From [35]. d From [113]. e From [79]. a

b

were the average of the of various integral hydrates existing at the concentration of water in nitrobenzene saturated by water (0.168 mol/L at 25°C). The hydrates were characterized by their formation constants Km=[X-(H2O)m]/([X-(H2O)m-1][H2O]). In interesting studies of Stoyanov et al., the composition of the organic extracts in 1,2-dichloroethane and the bound water were studied by IR-spectroscopy [116, 117]. The extraction was studied of H+ cobalt bis(dicarbollide) 1- and of the strontium cation with this anion in the absence and presence of three Sr synergists: PEG 400, 15-crown-5 and 18-crown-6. It was concluded that the hydrogen ion is hydrated in the organic phase, forming there a (H5O2.4H2O)+ particle. This contrasts with other extracts, e.g, in tri-n-butylphosphate, where enters the water core in the form of reverse nanomicelles [118]. In the presence of PEG 400, the hydrating water molecules are completely replaced by the six COC and two COH groups, exactly fitting to the coordination number of Sr2+ (n=8). Because of involvement of the ending COH groups of the polyethylene glycol, the lower efficacy of alkyl substituted PEGs in extraction of Sr2+ could be explained. The behavior of the two studied crown ethers, according to the authors of Ref. 116 differs considerably. In the complex the central cation is coordinated equally by all 10 COC groups of the two crown molecules. In however, the central ion coordinates with eight COC groups of two crown molecules, while one or two molecules are additionally coordinated to Sr2+, thus increasing the coordination number to 9 or 10. Moreover, it was found that H3O+ interacts more fully with 18–C–6 (by 6 oxygen atoms) than with 15–C–5 (with 5 oxygen atoms, due to one linear H bond and two bifurcated H bonds). These two effects led to much higher extractability of Sr2+ with 15–C–5 [116].

Copyright © 2004 by Marcel Dekker, Inc.

376

Rais

The hydration numbers found for nitrobenzene as a solvent were plotted against their respective ∆hydrG° values as shown in Fig. 6. Good correlation is – observed both for cationic and anionic data. The only exception of Cl might be an inaccurate value in [111], since previously Kenjo and Diamond found for this ion a hydration number nhyd=3.3 [119]. The latter value would lie on the straight line in Fig. 6. Generally, the hydration of ions in the organic phase seems to be one of the decisive factors, which determine the selectivity, especially of ions which are moderately to strongly hydrated in the organic phase. The information on hydration numbers should be viewed as subsidiary to actual ∆tG° values in concrete systems. 6.

Role of Dissolved Water in the Organic Solvent on Alkali Metal Cation Selectivity

For structurally similar solvents with comparable basicities, as are the nitrosolvents, the selectivity of extraction of alkali metal cations could be correlated as smooth monotonous curves if log K(Cs+/M+) were plotted against the solubility of water in the respective solvent [120]. This correlation applied for pure nitromethane, and water-saturated nitrobenzene, 2-nitropropane, nitroethane, and nitromethane solvents, in order of the increasing water solubility. With the newly obtained results plotted in Fig.5, considered importance of water solubility on the selectivity becomes more complicated. The curves for the transfers into nitromethane and o-nitro phenyl octyl ether are practically the same as in the Fig. 5. However, the water solubility in nitro-methane (~1.2 mol/L [120]) is much higher than for o-nitrophenyl octyl ether (0.046 mol/L [90]) and considering simply water solubility as the criterion, the selectivity for the latter solvent ought to be higher even than that for nitrobenzene. This may be explained by supposing that solvation in NPOE differs from that of simple nitrosolvents by letting the ether oxygen of NPOE partially enter the solvation sphere of the cation. Due to the higher basicity of ethers than nitrosolvents, the slope of the dependence for NPOE is then expected to be lower than for simple nitrosolvents as is the case. The dependences for dry NM and water-saturated NM, NB, and NPOE depicted in Fig. 5 display quite characteristic features; namely, they obey the normalizing linearity of ∆tG° (X, aq(or)→or(aq)) on ∆hydr G° for alkali metal ions with the exception of Li+ (see legend to Fig. 5). From this point of view, only the Li+ cation is considerably affected, which is reasonable due to its extensive hydration. The shift of the data point for Li+ against the linear functions appears to be regular in the series of selectivity, i.e., increasing with decreasing slope of the straight-line dependences. The hydration in other solvents than nitrobenzene was not systematically studied, but the data point for Li + and nitromethane gives n hyd=18 [110]. Excessive hydration leads to a decrease of selectivity as discussed above for M+, M2+, and M3+ cations (Section V.B).

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

377

7. General Affinity Scale for Organic Solvents The scale of affinity of anions toward the nitrobenzene, i.e., picrate - 2,4-dinitrophenolate [125], i.e., exactly in the order predicted by Table 2. It is noted that often the extraction of metal calixarene complexes with perchlorate anion is higher or equal to that with thiocyanate [126] in agreement with the data in Table 2. The use of the hydrophobicity/lipophilicity scale based on ∆tG°(X, aq→or) or ∆tG°(X, aq(or)→or(aq)) for dealing with the extraction

Copyright © 2004 by Marcel Dekker, Inc.

378

Rais

constants of calixarene-metal complexes was proposed [126], but no reference set of the values was suggested there.

VI. CONCLUSIONS The extraction of electrolytes is a particular type of extraction because the mechanism is not generally amenable to the usual log-log analysis. The dependences on the ligand concentration or on acidity of the aqueous phase often display maxima characteristic for this mechanism of extraction. In this review, the most common cases encountered in praxis were discussed and a simple procedure for the calculation of the curves with maxima was given. It is shown that, contrary to common belief, the mechanism of extraction of electrolytes applies in many situations in which a neutral ligand is used. This is so, even if no particular hydrophobic anion is used, the role of counter-ion being in this case played by a mineral acid anion present in the system. The experimental results of the extraction studies as well as electrochemical studies of transfers across ITIES deal very often with nitrobenzene as a solvent. Critically examined data of standard molar transfer Gibbs energies and hydration numbers for this solvent were collected here. The former data may be used more generally for predicting the behavior of new systems and ions, based on the proposed correlations in the review: (1) straight-line correlations of ∆tG°(X, aq(or)→or(aq)) values on the respective standard molar Gibbs energies of hydration for small inorganic ions, (2) linear dependences of ∆tG°(X, aq(or)→or(aq)) on the number of carbon atoms or properly chosen ionic radius of the ion for tetraalkylammonium and tetraphenyl derivatives, and (3) a general scale of hydrophobicity of anions based on ∆tG°(X, aq(NB)→NB(aq)). These enable the prediction of particular behaviors in extraction systems both with complete ion dissociation or association. The transfer data for nitrobenzene and small ions permitted an independent check of the underlying principles in the determination of both individual standard molar Gibbs energies of hydration and energies of transfer, justifying dealing with and using the tabulated values. This positive check consists in finding out that the experimentally not accessible values of ∆tG° of cations and anions display the same pattern of behavior as experimentally accessible values of hydration numbers of cations and anions in nitrobenzene namely straight line dependence on respective Gibbs energies of hydration of the ions.

SYMBOLS 1-

cobalt bis(dicarbollide) anion, [(1,2-C2B9H11)2-3-Co]Standard distribution potential of an ion X in the system of mutually saturated aqueous

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes and organic phases, [Eq. (24) and (25)],

[M+]aq, [M+]or cM

Didion DM DN EXTRIT

Iaq ISE ITIES

Kex(M, zB). Kex(MLz+, zB

the inner potentials of the denoted phases Equilibrium concentration of M+ in aqueous or organic phase Total (initial) concentration of M in the system referred to one of the phases when va/vo=1; e.g., if 0.06 M H+B- is dissolved in the organic phase and 3 M HNO3 is used in the aqueous phase, then cH=3+0.06= 3.06; used in the program EXTRIT Total (initial) concentration of M in the aqueous or organic phase Didion(X)=[Xj]or/[Xj]aq, distribution coefficient of a individual ion, [Eq. (20)] Distribution coefficient [Eq. (19)] Donor number of the solvent Simple computer program with double iteration, the second immersed in the first one, for calculating especially the curves with maxima Ionic strength of the aqueous phase Ion selective electrode Interface of Two Immiscible Electrolyte Solutions; refers to electrochemical studies performed at the actual interface of the two solutions in equilibrium or very near to equilibrium Association constant of M, zB in aqueous or organic phase [Eq. (5)] Association constant of ML, zB in aqueous or organic phase, [Eq. (12)] Extraction constant of Mz+, zB-, dissociated in the aqueous phase and associated M, zB ion pair in theorganic phase, [Eq. (7)] Extraction constant with bonding of cation by ligand L, [Eq. (10)] Extraction constant with bonding of cation by n molecules of ligand L, similar to Kex(MLz+, zB- [Eq. (14)]

Copyright © 2004 by Marcel Dekker, Inc.

379

380 Kex(Mz+), Kex(B-), Kex(Mz+, zB-) Kexch(MLz+/zH+). Kexch(Mz+/zH+)

kpotA/B LTGW r1 vaq/vor ∆tG°(X, aq→or) ∆tG°(X, aq(or)→or(aq))

∆tG°(Cs+]rel) ∆tG°(ne)

␦BS (w→n-octanol) ␦L ␦M, zB. ␴or,aq

Rais Individual extraction constant of ion Mz+ or B-, [Eq. (21)] Extraction constant of fully dissociated Mz+, zB-, [Eq. (4)] Extraction exchange constant of exchange of complexed cation MLz+ for zH+, [Eq. (13)] Extraction exchange constant of exchange of Mz+ for zH+, [Eq. (8)] Mixed constant of reaction defined by Eq. (18) Mixed constant of reaction defined by Eq. (17) “Theoretical selectivity coefficient” at ISE, dependent only on the properties of the solvent Program for analysis of the extraction curves developed by Vanura [36] and based originally on the LETAGROP program Ionic radius Volume ratio of aqueous to organic phases Half-wave potential at cyclic voltammetry at ITIES Standard molar Gibbs energy of transfer of an ion X from the aqueous phase into the organic phase, kJ/mol [Eq. (23)] Standard molar Gibbs energy of transfer on an ion X from the aqueous phase saturated with organic phase into the organic phase saturated with aqueous phase, kJ/mol Relative value of ∆tG° referred to Cs+ ion Non-electrostatic component of the overall ∆tG° Stability constant of formation of MLz+ cation in aqueous or organic phase, [Eq. (11)] Partition constant of neutral solute between water and n-octanol used as a parameter in drug research Distribution constant of the neutral ligand L between the aqueous and organic phases, [Eq. (9)] Distribution constant of rion-dissociated ion pair [Eq. (6)] Interfacial tension at the water/organic phase interface

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

381

ACKNOWLEDGMENTS The author thanks Prof. Y.Marcus for editing the text and for his valuable comments. Dr. S.Wilke (former Martin-Luther-Universität Halle-Wittenberg, BRD) supplied some of his data before publication. New results pertaining to the ion transfers between various two solvents in equilibrium and interrelation of cluster with solvation energetics were obtained in collaboration with Dr. T.Okada from NIAST, Tsukuba, Japan and were financially supported by Ministry of Education of Czech Republic, contract ME 485. Financial support for studies of new nonpolar solvents for cobalt bis(dicarbollide)s and for compilation of the newest results provided by Czech Grant Agency (Grant No. 104/ 01/0142) is appreciated. The studies of 137Cs extraction by crown ethers and calix crowns were supported by the Japanese Ministry of Education MEXT through the project ARTIST and the author wishes to express his thanks for this support.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Kyrs, M.; Pelcik, J.; Polansky, P. Coll. Czech. Chem. Comm. 1960, 25, 2642. Krtil, J.; Fojtik, M.; Kyrs, M. Coll. Czech. Chem. Comm. 1962, 27, 2069. Rais, J.; Podesva, S.; Kyrs, M. Anal. Chim. Acta. 1966, 36, 90–96. Rais, J. Coll. Czech. Chem. Comm. 1971, 36, 1665–1669. Rais, J.; Selucky, P. Coll. Czech. Chem. Comm. 1971, 36, 2766. Rais, J.; Selucky, P.; Drazanova, S. Radiochem. Radioanal Lett. 1970, 195, 4. Selucky, P.; Rais, J. Radiochem. Radioanal Lett. 1971, 6, 211. Rais, J; Selucky, P.; Drazanova, S. J. Inorg. Nucl. Chem. 1971, 33, 3087. Rais, J.; Selucky, P.; Kyrs, M. J. Inorg. Nucl. Chem. 1976, 38, 1376–1378. Rais, J.; Grüner, B. In Solvent Extraction and Ion Exchange, Marcus, Y. Gupta, C.K., Eds.; Marcel Dekker, 2004; Vol. 17. Rais, J.; Kyrs, M. Coll. Czech. Chem. Comm. 1969, 34, 949. Rais, J.; Kyrs, M. In Solvent Extraction Conference Gothenburg 1966; North Holland Publ. Co: Amsterdam, 1967, 595 p. Diamond, R.M. J. Phys. Chem. 1957, 61, 69–74. Diamond, R.M. J. Phys. Chem. 1957, 61, 75–81. Diamond, R.M.; Tuck, D.G. In Extraction of Inorganic Compounds into Organic Solvents in Progress in Inorganic Chemistry; Cotton, F.A., Ed.; Interscience Publishers: New York, 1960. Kuznetsov, V.I. Extrakcia II; Gosatomizdat: Moscow, 1962. in Russian. Stary, J.; Kyrs, M.; Marhol, M. Separation Methods in Radiochemistry; Academia: Praha, 1975. In Czech. Morisson, G.H.; Freiser, H. Anal Chem. 1964, 36, 93. Rais, J. Coll. Czech. Chem. Comm. 1971, 36, 3253–3262. Girault, H.H. http://dcwww.epfl.ch/cgi-bin/LE/DB/InterrDB.pl. Last accessed 22nd April 2003.

Copyright © 2004 by Marcel Dekker, Inc.

382 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

Rais Osakai, T.; Ebina, K. J. Phys. Chem. B. 1998, 102, 5691–5698. Osakai, T.; Ebina, K. In Liquid Interfaces in Chemical, Biological, and Pharmaceutical Applications; Volkov, A.G., Ed.; Marcel Dekker: New York, 2000. Marcus, Y. In Ion Solvation in Liquid—Liquid Interfaces, Theory and Methods; Volkov, A.G., Deamer, D.W., Eds.; CRC Press: New York, 1996. Marcus, Y. Ion Properties; Marcel Dekker: New York, 1997. Marcus, Y. Properties of Solvents; Wiley: Chichester, 1998. Gritzner, G. Electrochim Acta 1998, 44, 73–83. Marcus, Y. Electrochim Acta 1998, 44, 91–98. Moyer, B.A.; Sun, Y. In Ion Exchange and Solvent Extraction; Marinsky, J.A., Marcus, Y. Eds.; Marcel Dekker: New York, 1997; Vol. 13. Rais, J. Analytical Sciences 1998, 14, 293–297. Rais, J.; Kyrs, M.; Pivonková, M. J. Inorg. Nucl. Chem. 1968, 30, 611. Rais, J.; Kyrs, M. J. Inorg. Nucl. Chem. 1969, 31, 2903–3910. Bonnesen, P.V.; Delmau, L.H.; Haverlock, T.J.; Moyer, B.A. Report Oak Ridge National Laboratory, ORNL/TM-13704, December 1998. Vanura, P.; Jedinakova-Krizova, V. Solvent Extr and Ion Exchange 1992, 10, 847–859. Vanura, P.; Rais, J.; Selucky, P.; Kyrs, M. Coll. Czech. Chem. Comm. 1979, 44, 157. Vanura, P.; Makrlik, E.; Rais, J.; Kyrs, M. Coll. Czech. Chem. Comm. 1982, 47, 1444–1464. Vanura, P. Doctor of Science Thesis; Technical University of Prague: Prague, 2003. Sillén, L.G.; Warnqvist, B. Arkiv Kemi 1969, 31, 315. Vanura, P.; Benesova, M.; Makrlik, E.; Kyrs, M.; Rais, J. Coll. Czech. Chem. Comm. 1984, 49, 1367–1381. Rais, J.; Selucky, P.; Kadlecova, L. Report NRI, 7380-Ch, 1985. In Czech. Rais, J.; Alexova, J.; Okada, T. 9th International Seminar on Electroanalytical Chemistry; Changchun: China, 10–12 October 2003. Tachimori, S.; Yaita, T.; Rais, J. In Recent Progress in Actinides Separation Chemistry, Proceedings of the Workshop on Actinides Solution chemistry, WASC’94; Yoshida, Z., Kimura, T., Meguro, Y, Eds.; World Scientic: Singapore, 1997; 19–28. Rais, J.; Selucky, P.; Sistkova, N.V. International conference ATALANTE 2000, Avignon: France, 24–26 October 2000, poster No. 3.39. Yoshida, Y.; Matsui, M.; Shirai, O.; Maeda, K.; Kihara, S. Analyt. Chim Acta 3731998,, 213–225. Kielland, J. J. Am. Chem. Soc. 1937, 59, 1675. Pitzer, K.S. In Activity Coefficients in Electrolyte Solutions, 2nd Ed.; Pitzer, K.S., Ed.; CRC Press, Boca Raton, Fla., 75–151. Hildebrand, J.H.; Scott, R.L. The Solubility of Nonelectrolytes; 3rd Ed.; Reinhold Publishing Corp.: New York, 1950. Delmau, L.H.; Haverlock, T.J.; Levitskaia, T.G.; Sloop, F.V., Jr.; Moyer, B.A. Report of Oak Ridge National Laboratory, ORNL/TM-2001/267, December 2001. Koczorowski, Z.; Geblewicz, G. J. Electroanal. Chem. 1983, 152, 55–66. Kalidas, C.; Hefter, G.; Marcus, Y. Chem. Rev. 2000, 100, 819–852. Rais, J. Not published results. Marecek, V. Personal communication. Samec, Z.; Marecek, V.; Koryta, J.; Khalil, M.W. J. Electroanal. Chem. 1977, 83, 393.

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

383

Wilke, S.; Zerihun, T. J. Electroanal. Chem. 2001, 515, 52–60. Makrlik, E.; Bozek, F. Polish J. Chem. 1998, 72, 949–952. Conti, F.; Eisenman, G. Biophys. J. 1966, 6, 227. Scholer, R.; Simon, W. Helvetica Chim. Acta. 1972, 55, 1801–1809. Rais, J.; Mason, C.V.; Abney, K.D. Sep. Sci. Technology 1997, 32, 951–969. Langmaier, J.; Stejskalova, K.; Samec, Z. J. Electroanal. Chem. 2001, 496, 143–147. Sustacek, V.; Senkyr, J. J. Electroanal. Chem. 1990, 279, 31–41. Senkyr, J.; Kouril, K.J. Electroanal. Chem. 1984, 180, 383–393. Makrlik, E. Polish J. Chem. 1997, 71, 396–399. Baruzzi, A.M.; Wendt, H.J. Electroanal. Chem. 1990, 279, 19. Homolka, D.; Marecek, V.; Samec, Z.; Base, K.; Wendt, H. J. Electroanal. Chem. 1984, 163, 159. Takeda, Y.; Esaki, T.; Kudo, Y.; Matsuda, H. Bull. Chem. Soc. Jap. 1995, 63, 787–790. Su, Bin; Zhang, Shen; Guo, Jidong; Gan, Liangbing. Anal. Chem. 2002, 74, 373–378. Makrlik, E.; Rais, J.; Base, K.; Plesek, J.; Vanura, P. J. Radioanal. Nucl. Chem. Articles 1995, 198, 359–365. Wilke, S. J. Electroanal. Chem. 2001, 504, 184–194. Yoshida, Y.; Yoshida, Z.; Aoyagi, H.; Kitatsuji, Y.; Uehara, A.; Kihara, S. Anal Chim Acta 2002, 452, 149. Samec, Z.; Marecek, V.; Colombini, M.P. J. Electroanal Chem. 1988, 257, 147–154. Hundhammer, B.; Solomon, T.; Zerihun, T.; Abegaz, M.; Bekele, A.; Graichen, K. J. Electroanal Chem. 1994, 371, 1. Danil de Namor, A.F.; Hill, T. J. Chem. Soc. Fraraday Trans. 1983, 1 (79), 2713–2722. Hundhammer, B.; Solomon, T. J. Electroanal. Chem. 1983, 157, 19–26. Hundammer, B.; Solomon, T.; Alemu, H. J. Electroanal. Chem. 1983, 149, 179–183. Vanura, P. Czechoslovak J. Phys. 1999, 49 (Suppl. S1), 761–767. Marecek, V.; Samec, Z. Anal. Chim. Acta. 1983, 151, 265. Kitatsuji, Y.; Yoshida, Z.; Kudo, H.; Kihara, S. J. Electroanal. Chem. 2002, 520, 133–144. Makrlik, E.; Vanura, P. J. Electroanal. Chem 1990, 287, 159. Wilke, S.; Zerihun, T. Electrochimica Acta 1998, 44, 15–22. Podzimek, I.; Kyrs, M.; Rais, J. J. Inorg. Nucl. Chem. 1980, 42, 1481–1486. Podzimek, I.; Kyrs, M.; Rais, J.; Vanura, P. Polyhedron 1983, 2, 331–337. Scholz, F.; Komorsky-Lovric, S.; Lovric, M. Electrochem. Comm. 2000, 2, 112–118. Hundhammer, B.; Wilke, S. J. Electroanal. Chem. 1989, 266, 133–141. Wilke, S. J. Electroanal. Chem. 1991, 301, 67–75. Wandlowski, T.; Marecek, V.; Samec, Z. Electrochim Acta 1990, 35, 1173–1175. Makrlik, E. Coll. Czech. Chem. Comm. 1991, 56, 760–763. Rais, J.; Selucky, P.; Kyrs, M.; Kadlecova, L. J. Radioanal. Nucl. Chem. Articles 1994, 178, 27–39. Osakai, T.; Kakutani, T.; Nishiwaki, Y.; Senda, M. Bunseki Kagaku 1983, 32, E81– E84. Kakiuchi, T.; Noguchi, J.; Kotani, M.; Senda, M. J. Electroanal. Chem. Interfacial. Electrochem. 1990, 296, 517.

Copyright © 2004 by Marcel Dekker, Inc.

384 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122.

Rais Samec, Z. In Liquid-Liquid Interfaces, Theory and Methods; Volkov, A.G., Deamer, D.W., Eds.; CRC Press: New York, 1996. Samec, Z.; Langmaier, J.; Trojanek, A. J. Electroanal. Chem. 1996, 409, 1–7. Makrlik, E.; Halova, J.; Kyrs, M. Coll. Czech Chem. Comm. 1984, 49, 39–44. Makrlik, E. Colloids and Surfaces 1992, 68, 291–293. Rais, J.; Selucky, P. (in cooperation with M. Kyrs), Final report on contract No.34 PAK 93251017 for Rockwell International Corporation, Canoga Park, California, December 30, 1993. Marcus, Y. Ion Solvation; Wiley-Interscience: New York, 1985. Paulenova, A.; Svec, V.; Kopunec, R. J. Radioanal. Chem. Articles 1991, 150, 303–308. Rais, J.; Okada, T. J. Phys. Chem. A. 2000, 104, 7314–7323. Rais, J.; Alexova, J.; Okada, T. Anal. Chim. Acta. to be published. Rais, J.; Selucky, P.; Sistkova, N.V.; Alexova, J. Separation Science and Technology 1999, 34, 2865. Haugen, G.R.; Friedman, H.L. J. Phys. Chem. 1968, 72, 4549. Rais, J.; Okada, T. Phys. Chem. Chem. Phys. to be published. Vilcek, S. Diploma Work, NRI, 1974. in Slovak. Kyrs, M.; Selucky, P; Rais, J. Czech. Patent CS 219525, 4 December 1985. Krumgalz, B.S. J. Chem. Soc. Faraday I 1980, 76, 1275. Masadome, T.; Imato, T.; Wakida, S.; Higashi, K. Analytical Sciences 2000, 16, 383–389. Rais, J.; Plesek, J.; Kyrs, M.; Kadlecova, L. J. Radioanal Nucl. Chem. Articles 1991, 148, 349–357. Wang, H.; Yao, Y; Wang, E. J. Electroanal. Chem. 1994, 372, 21–25. Levins, R.J.; Ikeda, R.M. Anal. Chem. 1965, 37, 671. Rais, J.; Sebestova, E.; Selucky, P; Kyrs, M. J. Inorg. Nucl. Chem. 1976, 38, 1742. Iwachido, T. Bull. Chem. Soc. Jap. 1971, 44, 1835–1840. Skarda, V; Rais, J.; Kyrs, M. J. Inorg. Nucl. Chem. 1979, 41, 1443–1446. Osakai, T.; Ogata, A.; Ebina, K. J. Phys. Chem. B. 1997, 101, 8341. Iwachido, T.; Minami, M.; Kimura, M.; Sadakane, A.; Kawasaki, M.; Toei, K. Bull. Chem. Soc. Jap. 1980, 53, 703. Makrlik, E. Coll. Czech. Chem. Comm. 1992, 57, 289–295. Ogata, A.; Tsujino, Y; Osakai, T. Phys. Chem. Chem. Phys. 2000, 2, 247–251. Osakai, T.; Hoshino, M.; Izumi, M.; Kawakami, M.; Akasaka, K. J. Phys. Chem. B. 2000, 104, 12021–12027. Smirnov, I.V.; Stoyanov, E.S.; Vorob’eva, T.P. 14. Radiochemical conference; Marianske Lazne: Czech Republic, 14–19 April 2002. Varnek, A.; Wipff, G.; Famulari, A.; Raimondi, M.; Vorob’eva, T.; Stoyanov, E. J. Chem. Soc. Perkin. Trans. 2002, 2, 887–893. Stoyanov, E. J. Chem. Soc. Far. Trans. 1998, 94, 2803–2812. Kenjo, T.; Diamond, R.M. J. Inorg. Nucl. Chem. 1974, 36, 183. Selucky, P.; Rais, J.; Krtil,J. International Solvent Extraction Conference, The Hague 1971; paper 79, 14–17 p. Naganawa, H.; Suzuki, H.; Tachimori, S.; Nasu, A.; Sekine, T. Phys. Chem. Chem. Phys. 2001, 3, 2509–2517. Chmutova, M.K.; Litvina, M.N.; Nestorova, N.P.; Myasoedov, B.F.; Kabachnik, M.I. Solvent Extraction Ion Exchange 1992, 10, 439.

Copyright © 2004 by Marcel Dekker, Inc.

Principles of Extraction of Electrolytes

385

123. Kumar, A.; Mohapatra, P.K.; Pathak, P.N.; Manchanda, V.K. Talanta 1997, 45, 387–395. 124. Banerjee, S.; Mukhopadhyay, K.; Mukhopadhyay, B.; Lahiri, S. J. Radioanal. Nucl. Chem. 2002, 252, 157–160. 125. Katsuta, S.; Yamada, H.; Kudo, Y.; Takeda, Y. Anal. Chim. Acta. 2000, 416, 145–150. 126. Danil de Namor, A.F.; Cleverley, R.M.; Zapata-Ormachea, M.L. Chem. Rev. 1998, 98, 2495–2525.

Copyright © 2004 by Marcel Dekker, Inc.